Multitask prompted finetuning (MTF) has been shown to help large language
models generalize to new tasks in a zero-shot setting, but so far explorations
of MTF have focused on English data and models. We apply MTF to the pretrained
multilingual BLOOM and mT5 model families to produce finetuned variants called
BLOOMZ and mT0. We find finetuning large multilingual language models on
English tasks with English prompts allows for task generalization to
non-English languages that appear only in the pretraining corpus. Finetuning on
multilingual tasks with English prompts further improves performance on English
and non-English tasks leading to various state-of-the-art zero-shot results. We
also investigate finetuning on multilingual tasks with prompts that have been
machine-translated from English to match the language of each dataset. We find
training on these machine-translated prompts leads to better performance on
human-written prompts in the respective languages. Surprisingly, we find models
are capable of zero-shot generalization to tasks in languages they have never
intentionally seen. We conjecture that the models are learning higher-level
capabilities that are both task- and language-agnostic. In addition, we
introduce xP3, a composite of supervised datasets in 46 languages with English
and machine-translated prompts. Our code, datasets and models are publicly
available at https://github.com/bigscience-workshop/xmtf.
Figure 1: An overview of datasets in xP3. Datasets added to P3 in this work are marked bold. Yellow datasets are trained on. Green datasets are held out for evaluation.Figure 2: Language composition of xP3, ROOTS, and the corpus of mT5. All ROOTS and xP3 languages are depicted. The mT5 corpus covers additional languages that are not included in the graph.Figure 3: Comparison of dataset variants P3, xP3, and xP3mt on a sample from PAWS for P3 Zhang et al. (2019) and PAWS-X Yang et al. (2019) for xP3 and xP3mt. P3 pairs English datasets with English prompts, xP3 pairs multilingual datasets with English prompts and xP3mt pairs multilingual datasets with prompts machine-translated from English to match the dataset language. Expressions in curly brackets are replaced, e.g. for xP3mt the target shown as {{Choices[label]}}) becomes Sí.
Large language models pretrained on vast amounts of text show some capability of solving tasks expressed in natural language, even without explicit training on these tasks Brown et al. (2020). Finetuning on groups of language tasks has been shown to significantly boost this zero-shot task generalization of language models Wei et al. (2021); Sanh et al. (2022); Min et al. (2021). For example, Sanh et al. (2022) finetune on tasks like summarization and question answering leading to better performance on unseen tasks like natural language inference. Previous work has focused on multitask finetuning in the context of large English language models and tasks.
Multilingual large language models show the same zero-shot learning capabilities for both monolingual and crosslingual tasks Goyal et al. (2021a); Lin et al. (2021); Patel et al. (2022); Soltan et al. (2022)
. However, zero-shot performance tends to be significantly lower than finetuned performance. Thus, task-specific or language-specific transfer learning via finetuning remains the predominant practice
Devlin et al. (2018); Conneau et al. (2019). This is particularly challenging for low-resource languages or tasks with limited data available, such as writing a fable that teaches a specified moral. In the spirit of multitask finetuning, it would be desirable to improve the zero-shot task generalization of multilingual models to make them usable on tasks from low-resource languages without requiring further finetuning.
To address this goal, we focus on crosslingual multitask finetuning. Due to the difficulty of collecting supervised task data in low-resource languages, previous work typically aims to transfer capabilities learned from finetuning on English data, which can improve performance on non-English language tasks Wu and Dredze (2019); Chalkidis et al. (2021); Vu et al. (2022). We investigate whether English-only multitask finetuning also improves performance on non-English held-out tasks using the multilingual BLOOM BigScience Workshop (2022) and mT5 Xue et al. (2020) models. We find that after finetuning on the English-only multitask mixture used for T0 Sanh et al. (2022) (P3), performance on a diverse set of non-English held-out tasks increases.
To investigate whether multilingual task data can further improve performance, we extend P3 to xP3 by adding datasets from 46 different languages that cover tasks previously not present in P3 (such as translation and program synthesis). Finetuning on xP3 leads to even better zero-shot task generalization in both English and non-English compared to the P3-trained baseline. Models finetuned on xP3 perform best on English prompts, even for non-English samples. Hypothesizing that better performance could be attained by training on non-English prompts, we construct a variant of xP3 with machine-translated prompts called xP3mt. We find that finetuning on machine-translated prompts is enough to significantly increase performance on held-out tasks with non-English human-written prompts. However, reducing the number of English prompts in the finetuning also worsens English prompt performance on multilingual tasks.
Notably, we also find that models finetuned on xP3 generalize to held-out tasks in languages never intentionally seen during pretraining nor finetuning. We conduct a contamination analysis and find that only small amounts of these languages were included in the pretraining corpus. Thus, we hypothesize the models learn some language- and task-agnostic capabilities.
We publicly release all our datasets and models (URLs in Appendix §D).
2 Related work
2.1 Multitask learning
Multitask finetuning Sanh et al. (2022) (or instruction tuning Wei et al. (2021)) has emerged as a recipe for improving the zero-shot task generalization of large language models. Typically, these works define a task as a collection of datasets that require a certain set of skills. To inform large language models which task to perform given an input, a prompt is used to add natural language instructions to dataset instances Schick and Schütze (2020); Scao and Rush (2021). In this line of work, zero-shot task generalization refers to the ability to perform a held-out task based on prompted instructions alone. Our work builds on T0 (Sanh et al., 2022), a variant of T5 (Raffel et al., 2020) that underwent MTF and was subsequently shown to have strong zero-shot task generalization capabilities.
Increasing the number and diversity of finetuning tasks and datasets has been shown to increase model performance Min et al. (2021); Fries et al. (2022); Wang et al. (2022c); Scialom et al. (2022); Chung et al. (2022); Mishra et al. (2021b). PromptSource Bach et al. (2022) is a software application that provides a framework for developing and applying prompts. PromptSource was used to construct P3, the training dataset of T0. While most prior work has focused on using English prompts on English datasets, Wang et al. (2022b) trained both English and multilingual models on prompted datasets. Their multilingual model, called mTk-instruct, attains strong crosslingual performance. In contrast with Wang et al. (2022b), our sole focus is crosslingual zero-shot generalization. Therefore, we consider a wider variety of prompting settings and perform a more detailed evaluation of multilingual capabilities. Separately, Radford et al. (2019) find that accidental inclusion of non-English text gave the GPT-2 model a limited ability to process and generate non-English text. We similarly discover that our finetuned models can process text in languages not intentionally trained on.
2.2 Multilingual models
Many language models are pretrained on English data only. Multilingual pretrained language models Lample and Conneau (2019); Conneau et al. (2019); Fan et al. (2021) aim to enable processing a wide variety of non-English languages. Unlike monolingual models, multilingual models can also be used for crosslingual tasks, such as translation. For language generation, recent efforts have focused on two different model architectures based on the Transformer Vaswani et al. (2017). On the one hand, encoder-decoder transformers trained with a denoising objective such as mBART Liu et al. (2020) and mT5 Xue et al. (2020) learn to predict tokens masked out in the input sequence. Predicting masked tokens is only a pretraining task and these models are generally finetuned on downstream datasets before being used. On the other hand, decoder-only models pretrained on next token prediction such as mGPT Shliazhko et al. (2022), XGLM Lin et al. (2021) and BLOOM BigScience Workshop (2022) can be used to solve tasks expressed in natural language directly in a zero-shot or few-shot setting Brown et al. (2020). XGLM demonstrated competitive few-shot performance even when the model was prompted in a language different than the sample being processed. In particular, using English prompts for multilingual datasets provides better performance with XGLM than human-translating the English prompt to the dataset language.
In this work, we use the BLOOM models BigScience Workshop (2022); Scao et al. (2022), which were pretrained on the ROOTS corpus Laurençon et al. (2022) in 46 natural languages and 13 programming languages. We also finetune mT5 Xue et al. (2020) to compare encoder-decoder and decoder-only performance. mT5 is pretrained on a corpus sampled from mC4 covering 101 languages.
3 Finetuning data and models
To study crosslingual multitask prompted finetuning, we create xP3 by extending the P3 dataset collection with additional non-English tasks. We finetune both BLOOM and mT5 models on xP3. We refer to Appendix §D for public links to released models and datasets.
3.1 Finetuning data
We build on the P3 Sanh et al. (2022) task taxonomy and add 28 new multilingual datasets illustrated in Figure 1
. We define four task clusters previously not present in P3: translation, simplification, program synthesis, and miscellaneous code datasets. As 11% of BLOOM’s pretraining data is code, we add code datasets classified as program synthesis (text-to-code) or miscellaneous. The latter includes tasks such as estimating the computational complexity of a provided code snippet and generating a name for a given function. We extend the XWinograd dataset
Tikhonov and Ryabinin (2021) with winograd schemas from CLUE Xu et al. (2020) to increase its Chinese samples from 16 to 504. Similar to P3, a fraction of our prompts invert the task at hand. For example, a prompt may invert a closed-book QA sample by asking the model to generate a question given an answer.
With xP3 we aim to replicate the language distribution of the ROOTS corpus Laurençon et al. (2022) used to pretrain BLOOM. Thus, xP3 consists of the same 46 natural languages and code as ROOTS. ROOTS, xP3 and the mT5 corpus Xue et al. (2020) language distributions are visualized in Figure 2. 39% of xP3 data is English, slightly more than the 30% of English data in ROOTS. Various African languages such as Twi (tw) and Bambara (bm) form the tail of xP3’s language distribution. Many of them are not included in the mT5 pretraining corpus. In xP3, Twi and others are represented solely as a translation task using data from Flores-200 NLLB Team et al. (2022).
To study the importance of non-English prompts, we construct a machine-translated variant of xP3, xP3mt. We translate prompts of monolingual datasets into the respective dataset language. For example, for the Chinese dataset C3 Sun et al. (2020) prompts in xP3mt are in Chinese instead of English in xP3. For crosslingual datasets prompts remain in English (such as Wiki-Lingua, which involves producing a summary in one language based on text in another language). We use the Google Cloud API for machine translation111https://cloud.google.com/translate. Figure 3 compares the dataset variants we train on.
3.2 Models
Figure 4: Zero-shot multilingual task generalization with English prompts. BLOOM models have 176 billion parameters. Scores are the language average for each task. Appendix §B breaks down performance by language.
We use publicly available pretrained BLOOM models ranging from 560 million to 176 billion parameters. BLOOM models are large decoder-only language models pretrained for around 350 billion tokens with an architecture similar to GPT-3 Brown et al. (2020). We finetune the models for an additional 13 billion tokens with loss only being computed on target tokens. For example, given the input “Translate to English: Je t’aime.” and a space-separated target “I love you.”, the model is trained to predict only the targets. As targets vary in length from just one to hundreds of tokens, we downscale the loss of each token by the length of the target it belongs to. This ensures short targets (e.g. for multiple-choice QA) get the same weight as long targets (e.g. for translation). We skip samples longer than 2048 tokens and use packing to train efficiently on multiple samples at a time Kosec et al. (2021). We select the final checkpoint based on validation performance.
For mT5 models, we finetune using the T5X Roberts et al. (2022) framework on TPUs. mT5 uses the same encoder-decoder architecture, pretraining objective (masked language modeling), and pretraining length (1 trillion tokens) as T5 Raffel et al. (2020). For finetuning mT5, we follow the same procedure as described above for BLOOM, except that inputs are fed into the encoder and thus are not space-separated from targets.
We produce three core model variants available in different sizes:
BLOOMZ-P3 / mT0-P3: Models finetuned on the English-only P3.
BLOOMZ / mT0: Models finetuned on xP3, which consists of multilingual datasets with English prompts.
BLOOMZ-MT / mT0-MT: Models finetuned on xP3mt, which consists of multilingual datasets with English and machine-translated prompts.
We evaluate on three held-out tasks: coreference resolution, sentence completion and natural language inference (NLI) as depicted in Figure 1. We also evaluate on HumanEval due to its popularity for code evaluations Chen et al. (2021). For datasets that involve choosing the correct completion from several options, we follow prior work Sanh et al. (2022); Brown et al. (2020) and use rank classification: We compute the log-likelihood of each possible completion and select the highest scoring option. For each evaluation dataset, we select 5 prompts at random from PromptSource and use them for all language splits of the dataset. We report the median of the 5 prompts for results per language split. Thus, in constrast to XGLM Lin et al. (2021), we do not tune prompts based on performance on validation data. A selection of prompts can be found in Appendix §L. For generation evaluations we use lm-evaluation-harness Gao et al. (2021).
4 Results
Figure 5: Zero-shot task and language generalization using English prompts on tasks and languages not intentionally seen during pretraining nor finetuning. Language codes are ISO 639-1, except for JP (Japanese).
We first examine generalization to new tasks in languages included in finetuning in §4.1. Then, in §4.2, we look at language generalization: Can models generalize to tasks in languages that (a) they have only seen during pretraining and (b) they have never seen intentionally? In §4.3, we investigate performance on multilingual prompts and finetuning on xP3mt. Scaling laws are analyzed in §4.4. Finally, §4.5 looks at performance on generative tasks and §4.6 at the effect of language proportions on performance.
4.1 Task generalization
Previous work has shown that large language models finetuned on prompted multitask mixtures generalize to unseen tasks Zhong et al. (2021); Wei et al. (2021); Mishra et al. (2021b, a); Wang et al. (2022b). In Figure 4, we show that the same applies to multilingual models: Finetuned BLOOMZ and BLOOMZ-P3 models significantly improve over BLOOM and XGLM on held-out tasks. Despite an order of magnitude fewer parameters, mT0 (13 billion parameters) is ahead of BLOOMZ (176 billion parameters). We attribute this to the encoder-decoder architecture paired with a masked language modeling pretraining objective Wang et al. (2022a); Tay et al. (2022a) as well as the longer pretraining of mT5 Hoffmann et al. (2022); Su et al. (2022) (1 trillion tokens for mT5 vs. 366 billion for BLOOM). Despite also having gone through crosslingual multitask finetuning, mTk performs significantly worse than the same-sized mT0. We attribute this to our prompting style, which aims to replicate natural human communication. mTk is finetuned on more structured prompts with specific “Definition”, “Input” and “Output” fields. Similarly, Wang et al. (2022b) find that T0 performs worse than Tk on their prompts. We also find models finetuned on xP3 (BLOOMZ, mT0-13B) with 39% of English data outperform models finetuned on P3 (BLOOMZ-P3, mT0-13B-P3), which is 100% English (see Appendix §B). Even the fully English T0-11B model Sanh et al. (2022) is outperformed by our mT0-13B model. Ignoring embedding parameters these models have about the same size. This is likely due to xP3 adding additional tasks and prompts, which has been shown to help generalization Chung et al. (2022).
4.2 Language generalization
Here we add another layer of generalization: languages. Figure 4 already shows that finetuning on English data only (P3) leads to better performance on non-English data: For example, BLOOMZ-P3 improves by over 50% on multilingual sentence completion compared to BLOOM. Thus, zero-shot task performance in languages only seen during pretraining improves after finetuning on English. This has major practical benefits as it can be more difficult to collect data for low-resource languages.
Next, we investigate performance on languages the model has never intentionally seen. Due to the scale of large language model pretraining, it is difficult to label tasks or languages as strictly unseen. It is likely that the training data unintentionally includes small fractions of these languages (just as many tasks might appear “implicitly” in the pretraining corpus Sanh et al. (2022)). In Figure 5 we show that after multitask finetuning on xP3, the models can perform unseen tasks in languages that were not intentionally trained on. After probing the pretraining corpus in Appendix §C, we do find small amounts of these languages that were not intentionally included in the ROOTS corpus Laurençon et al. (2022). However, for XNLI, performance increases across all languages, many of which only show up in tiny fractions in our language contamination analysis, such as Thai with 0.006%. If we extrapolate this proportion to the entire ROOTS corpus, the BLOOM models would have seen a mere 20 million tokens of Thai during pretraining. One possibility is that better-than-random XNLI performance can be attained with little or no language understanding. In Appendix §G, we investigate Levenshtein distances of XNLI samples and find that there are meaningful differences across labels. Thus, sole inspection of characters without language understanding may be enough for better-than-random performance.
4.3 Multilingual prompting
Task
Prompt
Average accuracy
BLOOMZ
BLOOMZ-MT
mT0-13B
mT0-13B-MT
XNLI
EN
53.58
49.74
48.43
51.52
MT
37.87
42.03
39.83
42.64
HT
41.13
44.55
45.19
47.03
XCOPA
EN
75.5
75.75
84.45
81.6
MT
71.95
74.25
82.9
81.1
XStoryCloze
EN
84.42
84.07
82.52
82.58
MT
84.37
85.31
84.01
83.31
XWinograd
EN
60.07
59.15
70.49
73.24
MT
58.48
60.14
66.89
72.33
Table 1: Comparison between EN (English), MT (machine-translated) and HT (human-translated) prompts for 176B BLOOMZ and 13B mT0 models finetuned on either only English or English and machine-translated multilingual prompts (-MT).
Since all prompts in xP3 are in English (even for multilingual datasets), we created xP3mt, an extension with machine-translated prompts. To investigate performance on non-English prompts, we additionally human- and machine-translated the English prompts used for evaluation. In Table 1, we report performance when prompting in non-English languages. BLOOMZ performs much better on English than on non-English prompts. BLOOMZ-MT, which is finetuned on xP3mt, significantly improves on multilingual prompts. On XNLI, BLOOMZ-MT raises the average performance on human-translated prompts from 41.13 to 45.55. This comes at the cost of a reduction in its performance on English prompts, from 53.58 to 49.74. For mT0, the MT version provides similar performance gains on XNLI and XWinograd non-English prompts, while results on XCOPA and XStoryCloze are mixed. Similar to Lin et al. (2021), we also find that models perform better on human-translated prompts than machine-translated ones for XNLI.
4.4 Scaling
Figure 6: Aggregate performance vs. size. Transparent lines correspond to individual languages, while thick lines are average accuracy scores.
In Figure 4, the average performance of BLOOM is near the random baselines of 0.50 for Sentence Completion and Coreference Resolution and 0.33 for NLI. We think this is due to all of our experiments being zero-shot and using untuned prompts Perez et al. (2021a). We find in Figure 6 that even at 560M parameters, multitask finetuning improves zero-shot generalization. The gap between pretrained and multitask finetuned models grows significantly as parameters increase. Scaling up parameters benefits all languages evaluated.
4.5 Generation tasks
Figure 7: Validation performance during training on natural language understanding (NLU) and natural language generation (NLG) tasks. The former are scored using accuracy and the latter using BLEU Papineni et al. (2002). The NLG tasks measured are translation and summarization.
Pass@k
k=1
k=10
k=100
GPT-Neo 1.3B
4.79%
7.47%
16.30%
GPT-Neo 2.7B
6.41%
11.27%
21.37%
GPT-J 6B
11.62%
15.74%
27.74%
GPT-NeoX 20B
15.4%
25.6%
41.2%
Codex-300M
13.17%
20.37%
36.27%
Codex-679M
16.22%
25.7%
40.95%
Codex-2.5B
21.36%
35.42%
59.5%
Codex-12B
28.81%
46.81%
72.31%
BLOOM-560M
0.82%
3.02%
5.91%
BLOOM-1.1B
2.48%
5.93%
9.62%
BLOOM-1.7B
4.03%
7.45%
12.75%
BLOOM-3B
6.48%
11.35%
20.43%
BLOOM-7.1B
7.73%
17.38%
29.47%
BLOOM
15.52%
32.20%
55.45%
BLOOMZ-560M
2.18 %
4.11%
9.00%
BLOOMZ-1.1B
2.63%
6.22%
11.68%
BLOOMZ-1.7B
4.38%
8.73%
16.09%
BLOOMZ-3B
6.29%
11.94%
19.06%
BLOOMZ-7.1B
8.06%
15.03%
27.49%
BLOOMZ
12.06%
26.53%
48.44%
BLOOMZ-P3
6.13%
11.79%
18.73%
Table 2:
Code continuation on HumanEval. Non-BLOOM results come from prior work Chen et al. (2021); Fried et al. (2022). Codex is a language model finetuned on code, while the GPT models Black et al. (2021); Wang and Komatsuzaki (2021); Black et al. (2022) are trained on a mix of code and text like BLOOM.
In this section, we investigate the impact of multitask finetuning on generative tasks. In Figure 7, we plot validation performance throughout the training process. We find that while performance on natural language understanding tasks continues to increase, generative performance jumps initially and then decreases. Relatedly, in Table 2, we find that multitask finetuning does not improve performance on HumanEval Chen et al. (2021). Only for small models, such as BLOOM-560M vs. BLOOMZ-560M, there are meaningful performance gains. When no code data is included in finetuning (BLOOMZ-P3) performance decreases significantly. mT0 models, which have not been pretrained on code fail to solve any HumanEval problems (see full results in Appendix §K). Given a Python docstring, HumanEval requires models to complete a function. Inspecting generations reveals that the multitask finetuned models are biased towards short generations. In Appendix §E, we show example solutions from HumanEval and compute average length statistics. BLOOMZ tries to solve problems with 70% fewer characters than BLOOM. One possible reason for this is that a majority of samples seen during multitask finetuning are only single sentences, so finetuned models learn to produce short answers. This could be causing the decreasing performance on generative tasks, which require longer answers than natural language understanding tasks. To force longer generations at inference time, we find it beneficial to enforce a minimum number of tokens during which the end-of-sequence token is ignored. We provide qualitative examples of forcing a minimum number of tokens in Appendix §F.
4.6 Effect of language proportions
Figure 8: Performance across languages by size in the BLOOM pretraining corpus, ROOTS.
In Figure 8, we find that finetuned BLOOM models perform better on languages seen extensively during pretraining. As the language distribution in the finetuning dataset, xP3, closely follows that of pretraining, these languages are also seen most frequently during finetuning. Specifically, XCOPA and XNLI show significantly better performance on these high-resource languages, such as English, Spanish or French, which all make up more than 10% of pretraining individually. The trend is less consistent for XWinograd. This may be caused by the fact that XWinograd language subsets are not translations of each other and have a significantly different number of samples. Thus, some language subsets of XWinograd may be inherently more difficult than others.
5 Conclusion
In this work we investigated crosslingual multitask finetuning. We developed xP3, a corpus consisting of tasks in 46 languages. Further, we have extended xP3 to xP3mt with machine-translated prompts. We have finetuned pretrained BLOOM and mT5 models on the newly created corpora as well as the English-only P3 corpus to produce BLOOMZ and mT0 models.
We found that English-only finetuning suffices for a multilingual pretrained large language model to generalize to tasks in other pretrained languages. However, finetuning on multiple languages using xP3 provided even better performance. We have further observed finetuned models to be capable of generalization to new tasks in languages they have never intentionally seen. We investigated multilingual prompting and found performance after finetuning on English prompts only to be poor. However, finetuning on a corpus with machine-translated prompts (xP3mt) lead to significantly better performance on human-written non-English prompts. Comparing models from 560 million up to 176 billion parameters revealed that the performance gap between only pretraining and finetuning widens as parameters increase. Lastly, we found multitask finetuning on billions of short targets biases models to produce short answers, which can hurt performance on generative tasks.
To contribute to future progress on improving zero-shot generalization, we release all datasets and models introduced in this work.
Black et al. (2021)
Sid Black, Leo Gao, Phil Wang, Connor Leahy, and Stella Biderman. 2021.
Gpt-neo: Large scale autoregressive language modeling with
mesh-tensorflow.
If you use this software, please cite it using these metadata,
58.
Bommasani et al. (2021)
Rishi Bommasani, Drew A Hudson, Ehsan Adeli, Russ Altman, Simran Arora, Sydney
von Arx, Michael S Bernstein, Jeannette Bohg, Antoine Bosselut, Emma
Brunskill, et al. 2021.
On the opportunities and risks of foundation models.
arXiv preprint arXiv:2108.07258.
Brown et al. (2020)
Tom Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared D Kaplan, Prafulla
Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell,
Sandhini Agarwal, Ariel Herbert-Voss, Gretchen Krueger, Tom Henighan, Rewon
Child, Aditya Ramesh, Daniel Ziegler, Jeffrey Wu, Clemens Winter, Chris
Hesse, Mark Chen, Eric Sigler, Mateusz Litwin, Scott Gray, Benjamin Chess,
Jack Clark, Christopher Berner, Sam McCandlish, Alec Radford, Ilya Sutskever,
and Dario Amodei. 2020.
Language models are few-shot learners.
In Advances in Neural Information Processing Systems,
volume 33, pages 1877–1901. Curran Associates, Inc.
Chung et al. (2022)
Hyung Won Chung, Le Hou, Shayne Longpre, Barret Zoph, Yi Tay, William Fedus,
Eric Li, Xuezhi Wang, Mostafa Dehghani, Siddhartha Brahma, Albert Webson,
Shixiang Shane Gu, Zhuyun Dai, Mirac Suzgun, Xinyun Chen, Aakanksha
Chowdhery, Sharan Narang, Gaurav Mishra, Adams Yu, Vincent Zhao, Yanping
Huang, Andrew Dai, Hongkun Yu, Slav Petrov, Ed H. Chi, Jeff Dean, Jacob
Devlin, Adam Roberts, Denny Zhou, Quoc V. Le, and Jason Wei. 2022.
Scaling
instruction-finetuned language models.
arXiv preprint arXiv:2210.11416.
Dhole et al. (2021)
Kaustubh D Dhole, Varun Gangal, Sebastian Gehrmann, Aadesh Gupta, Zhenhao Li,
Saad Mahamood, Abinaya Mahendiran, Simon Mille, Ashish Srivastava, Samson
Tan, et al. 2021.
Nl-augmenter: A framework for task-sensitive natural language
augmentation.
arXiv preprint arXiv:2112.02721.
Ding et al. (2022)
Ning Ding, Yujia Qin, Guang Yang, Fuchao Wei, Zonghan Yang, Yusheng Su,
Shengding Hu, Yulin Chen, Chi-Min Chan, Weize Chen, et al. 2022.
Delta tuning: A comprehensive study of parameter efficient methods
for pre-trained language models.
arXiv preprint arXiv:2203.06904.
Fan et al. (2021)
Angela Fan, Shruti Bhosale, Holger Schwenk, Zhiyi Ma, Ahmed El-Kishky,
Siddharth Goyal, Mandeep Baines, Onur Celebi, Guillaume Wenzek, Vishrav
Chaudhary, et al. 2021.
Beyond english-centric multilingual machine translation.
J. Mach. Learn. Res., 22(107):1–48.
Fries et al. (2022)
Jason Alan Fries, Leon Weber, Natasha Seelam, Gabriel Altay, Debajyoti Datta,
Samuele Garda, Myungsun Kang, Ruisi Su, Wojciech Kusa, Samuel Cahyawijaya,
et al. 2022.
Bigbio: A framework for data-centric biomedical natural language
processing.
arXiv preprint arXiv:2206.15076.
Gao et al. (2021)
Leo Gao, Jonathan Tow, Stella Biderman, Sid Black, Anthony DiPofi, Charles
Foster, Laurence Golding, Jeffrey Hsu, Kyle McDonell, Niklas Muennighoff,
Jason Phang, Laria Reynolds, Eric Tang, Anish Thite, Ben Wang, Kevin Wang,
and Andy Zou. 2021.
A framework for
few-shot language model evaluation.
Goyal et al. (2021b)
Naman Goyal, Cynthia Gao, Vishrav Chaudhary, Peng-Jen Chen, Guillaume Wenzek,
Da Ju, Sanjana Krishnan, Marc’Aurelio Ranzato, Francisco Guzm’an, and
Angela Fan. 2021b.
The flores-101 evaluation benchmark for low-resource and multilingual
machine translation.
Guzm’an et al. (2019)
Francisco Guzm’an, Peng-Jen Chen, Myle Ott, Juan Pino, Guillaume Lample,
Philipp Koehn, Vishrav Chaudhary, and Marc’Aurelio Ranzato. 2019.
Two new evaluation datasets for low-resource machine translation:
Nepali-english and sinhala-english.
Hasan et al. (2021)
Tahmid Hasan, Abhik Bhattacharjee, Md. Saiful Islam, Kazi Mubasshir, Yuan-Fang
Li, Yong-Bin Kang, M. Sohel Rahman, and Rifat Shahriyar. 2021.
XL-sum:
Large-scale multilingual abstractive summarization for 44 languages.
In Findings of the Association for Computational Linguistics:
ACL-IJCNLP 2021, pages 4693–4703, Online. Association for Computational
Linguistics.
Hellendoorn et al. (2020)
Vincent J. Hellendoorn, Charles Sutton, Rishabh Singh, Petros Maniatis, and
David Bieber. 2020.
Global relational
models of source code.
In 8th International Conference on Learning Representations,
ICLR 2020, Addis Ababa, Ethiopia, April 26-30, 2020. OpenReview.net.
Hendrycks et al. (2021)
Dan Hendrycks, Steven Basart, Saurav Kadavath, Mantas Mazeika, Akul Arora,
Ethan Guo, Collin Burns, Samir Puranik, Horace He, Dawn Song, and Jacob
Steinhardt. 2021.
Measuring coding challenge competence with apps.
NeurIPS.
Kim et al. (2021)
Joongwon Kim, Mounica Maddela, Reno Kriz, Wei Xu, and Chris Callison-Burch.
2021.
BiSECT:
Learning to split and rephrase sentences with bitexts.
In Proceedings of the 2021 Conference on Empirical Methods in
Natural Language Processing, pages 6193–6209, Online and Punta Cana,
Dominican Republic. Association for Computational Linguistics.
Lample and Conneau (2019)
Guillaume Lample and Alexis Conneau. 2019.
Cross-lingual
language model pretraining.
Laurençon et al. (2022)
Hugo Laurençon, Lucile Saulnier, Thomas Wang, Christopher Akiki,
Albert Villanova del Moral, Teven Le Scao, Leandro Von Werra, Chenghao Mou,
Eduardo González Ponferrada, Huu Nguyen, et al. 2022.
The bigscience roots corpus: A 1.6 tb composite multilingual dataset.
In Thirty-sixth Conference on Neural Information Processing
Systems Datasets and Benchmarks Track.
Levenshtein et al. (1966)
Vladimir I Levenshtein et al. 1966.
Binary codes capable of correcting deletions, insertions, and
reversals.
In Soviet physics doklady, volume 10, pages 707–710. Soviet
Union.
Muennighoff et al. (2022)
Niklas Muennighoff, Nouamane Tazi, Loïc Magne, and Nils Reimers. 2022.
Mteb: Massive text
embedding benchmark.
arXiv preprint arXiv:2210.07316.
NLLB Team et al. (2022)
NLLB Team, Marta R. Costa-jussà, James Cross, Onur Çelebi, Maha Elbayad,
Kenneth Heafield, Kevin Heffernan, Elahe Kalbassi, Janice Lam, Daniel Licht,
Jean Maillard, Anna Sun, Skyler Wang, Guillaume Wenzek, Al Youngblood, Bapi
Akula, Loic Barrault, Gabriel Mejia Gonzalez, Prangthip Hansanti, John
Hoffman, Semarley Jarrett, Kaushik Ram Sadagopan, Dirk Rowe, Shannon Spruit,
Chau Tran, Pierre Andrews, Necip Fazil Ayan, Shruti Bhosale, Sergey Edunov,
Angela Fan, Cynthia Gao, Vedanuj Goswami, Francisco Guzmán, Philipp Koehn,
Alexandre Mourachko, Christophe Ropers, Safiyyah Saleem, Holger Schwenk, and
Jeff Wang. 2022.
No language left
behind: Scaling human-centered machine translation.
arXiv preprint 2207.04672.
Papineni et al. (2002)
Kishore Papineni, Salim Roukos, Todd Ward, and Wei-Jing Zhu. 2002.
Bleu: a method for automatic evaluation of machine translation.
In Proceedings of the 40th annual meeting of the Association
for Computational Linguistics, pages 311–318.
Perez et al. (2021a)
Ethan Perez, Douwe Kiela, and Kyunghyun Cho. 2021a.
True few-shot learning with language models.
Advances in Neural Information Processing Systems,
34:11054–11070.
Ponti et al. (2020)
Edoardo M. Ponti, Goran Glavas, Olga Majewska, Qianchu Liu, Ivan Vuli’c, and
Anna Korhonen. 2020.
XCOPA: A
multilingual dataset for causal commonsense reasoning.
arXiv preprint.
Radford et al. (2019)
Alec Radford, Jeffrey Wu, Rewon Child, David Luan, Dario Amodei, Ilya
Sutskever, et al. 2019.
Language models are unsupervised multitask learners.
OpenAI blog, 1(8):9.
Rae et al. (2021)
Jack W Rae, Sebastian Borgeaud, Trevor Cai, Katie Millican, Jordan Hoffmann,
Francis Song, John Aslanides, Sarah Henderson, Roman Ring, Susannah Young,
et al. 2021.
Scaling language models: Methods, analysis & insights from training
gopher.
arXiv preprint arXiv:2112.11446.
Raffel et al. (2020)
Colin Raffel, Noam Shazeer, Adam Roberts, Katherine Lee, Sharan Narang, Michael
Matena, Yanqi Zhou, Wei Li, Peter J Liu, et al. 2020.
Exploring the limits of transfer learning with a unified text-to-text
transformer.
J. Mach. Learn. Res., 21(140):1–67.
Raganato et al. (2020)
Alessandro Raganato, Tommaso Pasini, Jose Camacho-Collados, and Mohammad Taher
Pilehvar. 2020.
Xl-wic: A multilingual benchmark for evaluating semantic
contextualization.
In Proceedings of the 2020 Conference on Empirical Methods in
Natural Language Processing (EMNLP), pages 7193–7206.
Roemmele et al. (2011)
Melissa Roemmele, Cosmin Adrian Bejan, and Andrew S Gordon. 2011.
Choice
of plausible alternatives: An evaluation of commonsense causal reasoning.
In 2011 AAAI Spring Symposium Series.
Sanh et al. (2022)
Victor Sanh, Albert Webson, Colin Raffel, Stephen Bach, Lintang Sutawika, Zaid
Alyafeai, Antoine Chaffin, Arnaud Stiegler, Teven Le Scao, Arun Raja, et al.
2022.
Multitask prompted training enables zero-shot task generalization.
In The Tenth International Conference on Learning
Representations.
Soltan et al. (2022)
Saleh Soltan, Shankar Ananthakrishnan, Jack FitzGerald, Rahul Gupta, Wael
Hamza, Haidar Khan, Charith Peris, Stephen Rawls, Andy Rosenbaum, Anna
Rumshisky, Chandana Satya Prakash, Mukund Sridhar, Fabian Triefenbach, Apurv
Verma, Gokhan Tur, and Prem Natarajan. 2022.
Alexatm 20b:
Few-shot learning using a large-scale multilingual seq2seq model.
Srivastava et al. (2022)
Aarohi Srivastava, Abhinav Rastogi, Abhishek Rao, Abu Awal Md Shoeb, Abubakar
Abid, Adam Fisch, Adam R Brown, Adam Santoro, Aditya Gupta, Adrià
Garriga-Alonso, et al. 2022.
Beyond the imitation game: Quantifying and extrapolating the
capabilities of language models.
arXiv preprint arXiv:2206.04615.
Sun et al. (2020)
Kai Sun, Dian Yu, Dong Yu, and Claire Cardie. 2020.
Investigating prior knowledge for challenging chinese machine reading
comprehension.
Trans. Assoc. Comput. Linguistics, 8:141–155.
Suzgun et al. (2022)
Mirac Suzgun, Nathan Scales, Nathanael Schärli, Sebastian Gehrmann, Yi Tay,
Hyung Won Chung, Aakanksha Chowdhery, Quoc V Le, Ed H Chi, Denny Zhou, et al.
2022.
Challenging big-bench tasks and whether chain-of-thought can solve
them.
arXiv preprint arXiv:2210.09261.
Tay et al. (2022a)
Yi Tay, Mostafa Dehghani, Vinh Q Tran, Xavier Garcia, Dara Bahri, Tal Schuster,
Huaixiu Steven Zheng, Neil Houlsby, and Donald Metzler. 2022a.
Unifying language learning paradigms.
arXiv preprint arXiv:2205.05131.
Tiedemann (2020)
J”org Tiedemann. 2020.
The Tatoeba
Translation Challenge – Realistic data sets for low resource and
multilingual MT.
In Proceedings of the Fifth Conference on Machine Translation,
pages 1174–1182. Association for Computational Linguistics.
Vaswani et al. (2017)
Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones,
Aidan N Gomez, Łukasz Kaiser, and Illia Polosukhin. 2017.
Attention is all you need.
Advances in neural information processing systems, 30.
Williams et al. (2018)
Adina Williams, Nikita Nangia, and Samuel Bowman. 2018.
A broad-coverage
challenge corpus for sentence understanding through inference.
In Proceedings of the 2018 Conference of the North American
Chapter of the Association for Computational Linguistics: Human Language
Technologies, Volume 1 (Long Papers), pages 1112–1122. Association for
Computational Linguistics.
Wu and Dredze (2019)
Shijie Wu and Mark Dredze. 2019.
Beto, bentz, becas: The
surprising cross-lingual effectiveness of BERT.
In Proceedings of the 2019 Conference on Empirical Methods in
Natural Language Processing and the 9th International Joint Conference on
Natural Language Processing (EMNLP-IJCNLP), pages 833–844, Hong Kong,
China. Association for Computational Linguistics.
Yang et al. (2019)
Yinfei Yang, Yuan Zhang, Chris Tar, and Jason Baldridge. 2019.
PAWS-X: A Cross-lingual Adversarial Dataset for Paraphrase
Identification.
In Proc. of EMNLP.
This research was conducted under the BigScience project for open research, a year-long initiative targeting the study of large models and datasets. The goal of the project is to research language models in a public environment. The project has hundreds of researchers from more than 50 countries and over 250 institutions. The BigScience project was initiated by Thomas Wolf at Hugging Face, and this collaboration would not have been possible without his effort. In the following, we list contributions made to this work.
Niklas Muennighoff evaluated all models, created xP3 and wrote most of the paper.
Niklas Muennighoff, Thomas Wang, Lintang Sutawika, Adam Roberts and Hailey Schoelkopf wrote the training and evaluation code.
Niklas Muennighoff and Adam Roberts trained the models.
Niklas Muennighoff, Teven Le Scao, Hailey Schoelkopf, Zheng-Xin Yong, Thomas Wang, Khalid Almubarak, Alham Fikri Aji, M Saiful Bari and Zaid Alyafeai contributed prompts or datasets.
Lintang Sutawika, Stella Biderman, Zheng-Xin Yong, Khalid Almubarak, M Saiful Bari and Albert Webson initiated the project.
Sheng Shen conducted the contamination analysis.
Samuel Albanie wrote the prompt appendix.
Thomas Wang and Zheng-Xin Yong converted checkpoints.
Colin Raffel, Thomas Wang, Teven Le Scao, M Saiful Bari, Edward Raff and Dragomir Radev advised the project.
Niklas Muennighoff, Lintang Sutawika, Teven Le Scao, Colin Raffel, Stella Biderman, Alham Fikri Aji, Adam Roberts, Samuel Albanie, Sheng Shen, M Saiful Bari, Albert Webson, Xiangru Tang, Dragomir Radev and Edward Raff contributed to the paper.
Appendix B Task generalization breakdown
In Figure 9, we compare performance on English held-out tasks. We find that (a) finetuning on xP3 outperforms P3 (b) multilingual mT0 is stronger than monolingual T0 on English. We conjecture that both improvements come from xP3 having more prompts and datasets than P3 Chung et al. (2022).
Figure 9: Zero-shot English task generalization. Each dot represents performance on one English evaluation prompt.
In Figure 10, we visualize task generalization to multilingual datasets. The same data is aggregated in Figure 4. Performance by prompt varies substantially highlighting that prompt engineering may still be necessary after MTF. We also find that mT0 consistently outperforms BLOOMZ on Swahili (SW), possibly due to it being a larger part of its pretraining corpus (see Figure 2 and §4.6).
Figure 10: Zero-shot multilingual task generalization on languages seen during pretraining and finetuning. Each dot represents performance on one English evaluation prompt.
Appendix C ROOTS language contamination
Figure 11: Language composition of ROOTS-IDENTIFY-1%, ROOTS-1% and the mT5 corpus. All mT5 languages are depicted. ROOTS-1% is a random 1% sample of ROOTS with its assigned meta-languages. ROOTS-IDENTIFY-1% are the actual languages in ROOTS-1% re-identified using cld3.
While the BLOOM ROOTS corpus Laurençon et al. (2022) was collected from 46 natural languages and 13 programming languages, we find that sentences from the same document do not always belong to the collected (meta) language. Some sentences use languages like Russian or Japanese that were not the intentionally collected parts. This “language contamination” may stem from “code-mixing” or different languages being used in code comments. To investigate the extent of contamination, we randomly sample 1% of the documents from ROOTS for a total of 51M documents. For each document, we use cld3222https://github.com/google/cld3 Xue et al. (2020) to identify the languages used in each sentence and compare them with the meta language of the document. We summarize our results in Figure 11. It shows that ROOTS contains unintentionally collected languages, such as Burmese (my: 0.00003%), Thai (th: 0.006%), Turkish (tr: 0.03%), Greek (el: 0.03%), Russian (ru: 0.03%), Bulgarian (bg: 0.05%), Estonian (et: 0.06%), Haitian (ht: 0.12%), German (de: 0.21%), Italian (it: 0.28%) and Japanese (ja: 0.54%). These “unseen” languages only have small sentence proportions in our subsample compared to English (en: 46.23%), French (fr: 15.73%) and Spanish (es: 13.38%). Yet, they may help the language generalization of BLOOMZ models described in §4.2. Japanese is mostly mixed in the meta English documents (47%), meta Code documents (8%) and meta Chinese documents (5%). Meanwhile, Russian is mostly mixed in the meta English documents (52%), meta Code documents (19%) and meta French documents (11%).
Appendix D Artifacts
Table 3 lists all artifacts used or released in this work.
Artifact
Explanation
Public link
ROOTS
Multilingual pretraining corpus of BLOOM
https://huggingface.co/bigscience-data
mC4
Multilingual pretraining corpus used for mT5
https://huggingface.co/datasets/mc4
P3
Multitask finetuning dataset with English data & English prompts
https://huggingface.co/datasets/bigscience/P3
xP3
Multitask finetuning dataset with multilingual data & English prompts
https://huggingface.co/datasets/bigscience/xP3
xP3all
Same as xP3 with held-out evaluation sets
https://huggingface.co/datasets/bigscience/xP3all
xP3mt
Same as xP3 with English & multilingual machine-translated prompts
https://huggingface.co/datasets/bigscience/xP3mt
xP3megds
Processed version of xP3 for easy usage with Megatron-DeepSpeed
300M parameter model pretrained on a sampled version of mC4
https://huggingface.co/google/mt5-small
mT5-580M
580M parameter model pretrained on a sampled version of mC4
https://huggingface.co/google/mt5-base
mT5-1.2B
1.2B parameter model pretrained on a sampled version of mC4
https://huggingface.co/google/mt5-large
mT5-3.7B
3.7B parameter model pretrained on a sampled version of mC4
https://huggingface.co/google/mt5-xl
mT5-13B
13B parameter model pretrained on a sampled version of mC4
https://huggingface.co/google/mt5-xxl
mT0-300M
300M parameter model finetuned on xP3
https://huggingface.co/bigscience/mt0-small
mT0-580M
580M parameter model finetuned on xP3
https://huggingface.co/bigscience/mt0-base
mT0-1.2B
1.2B parameter model finetuned on xP3
https://huggingface.co/bigscience/mt0-large
mT0-3.7B
3.7B parameter model finetuned on xP3
https://huggingface.co/bigscience/mt0-xl
mT0-13B
13B parameter model finetuned on xP3
https://huggingface.co/bigscience/mt0-xxl
mT0-13B-MT
13B parameter model finetuned on xP3mt
https://huggingface.co/bigscience/mt0-xxl-mt
mT0-13B-P3
13B parameter model finetuned on P3
https://huggingface.co/bigscience/mt0-xxl-p3
Table 3: Links to all models & datasets used as part of this work. BLOOMZ models have an additional repository containing the final optimizer states for training with Megatron-Deepspeed that can be found by appending “-optimizer-states” to the respective URL.
Appendix E Code generations
Table 4 provides statistics on code generations and code data. We find that BLOOM generates on average 70% more characters and 17x more comments than BLOOMZ for a given problem from HumanEval. Figure 12 compares an example solution from BLOOM and BLOOMZ. While both solutions are correct, BLOOMZ is biased towards short and concise answers.
(a) BLOOM
(b) BLOOMZ
Figure 12: Code generations of BLOOM and BLOOMZ on HumanEval. The model is prompted to generate after the final """. The generation is stopped after an end-of-sequence token or a return statement followed by a newline.
Data (→)
HumanEval generations
Targets of xP3
BLOOM
BLOOMZ
code datasets
Average characters
247
144
531
Average Python comments (#)
0.69
0.04
0.85
Table 4: Number of characters and comments for generations and targets in the finetuning corpus.
Appendix F Qualitative examples
(a) English prompt
(b) Non-English prompt
Figure 13:
Greedy generations for sentiment analysis, a task trained on. BLOOMZ and mT0-13B have not been trained on non-English prompts, but are still able to handle them. BLOOMZ, however, answers in English. The review is a five star review of Star Wars Episode IV.
(a) English prompt
(b) Non-English prompt
Figure 14: Greedy generations for zero-shot query expansion, a task not trained on. The models sometimes fail to output at least five terms as requested in the prompt.
(a) English prompt
(b) English prompt
Figure 15: Greedy generations on question answering, a task trained on. Left: Specifying the language in the prompt is an effective way to force the output language. Right:
Specifying a minimum token length as a generation hyperparameter is an effective way to force long generations. The output of BLOOM is shortened (marked with
...).
(a) English prompt
(b) English prompt
Figure 16: Non-greedy fable generations given a moral, a task not trained on. The generations are cherry-picked from 16 outputs with no minimum length, a temperature of 0.9 and top k of 40. Left: BLOOMZ generates an interesting fable with the desired moral. mT0 is significantly worse at writing stories likely due to its different pretraining objective. Right: BLOOMZ does not seem to understand the moral correctly.
Appendix G XNLI Levenshtein distances
To investigate whether XNLI can be solved without any language understanding, we compute Levenshtein distances Levenshtein et al. (1966) between premise and hypothesis and average them by the XNLI label. In Table 5 we find that the distances are smallest between entailment pairs and largest between neutral pairs. This is intuitive as entailment pairs generally need to cover similar content. Contradiction pairs still need to cover similar content but differ in at least one major way. Meanwhile for neutral pairs hypothesis and premise may be about completely different topics. This highlights that XNLI can be solved to some degree by solely comparing the similarity of characters across premise and hypothesis.
Label (→)
Entailment
Neutral
Contradiction
Language (↓)
Thai (th)
79.08
82.64
81.52
Turkish (tr)
76.93
80.59
80.24
Greek (el)
90.90
95.10
93.93
Table 5: Levenshtein distances between hypothesis and premise averaged across samples from different XNLI labels. Each label has 830 samples per language subset.
Appendix H Multilingual prompting in unseen languages
Table 6 shows aggregate performances on languages not intentionally seen during pretraining nor finetuning for BLOOMZ and only seen during pretraining for mT0. For BLOOMZ, performance drops significantly when translating the prompts to the respective unseen languages. Further, BLOOMZ-MT loses its edge over BLOOMZ as it has not been finetuned on prompts in these languages. For mT0 differences are less significant.
Task
Prompt
Average accuracy
BLOOMZ
BLOOMZ-MT
mT0-13B
mT0-13B-MT
XNLI
EN
45.65
43.2
48.52
51.33
MT
36.48
35.67
41.86
39.78
XCOPA
EN
54.27
53.67
72.67
71.6
MT
53.2
53.0
71.57
70.87
XStoryCloze
EN
61.59
61.36
79.31
80.13
MT
60.5
59.91
80.21
80.28
XWinograd
EN
55.98
54.54
70.81
72.0
MT
53.11
52.46
67.86
70.45
Table 6: Comparison between EN (English), MT (machine-translated) and HT (human-translated) prompts for 176B BLOOMZ and 13B mT0 models finetuned on either only English or English and machine-translated multilingual prompts (-MT). For BLOOMZ the evaluation languages averaged are never intentionally seen, such as Japanese and Russian for XWinograd (see Figure 5). For mT0 the evaluation languages are only seen during pretraining.
Appendix I Ideas that did not work
We list several experiments that did not improve over baseline results:
Non-causal
In a non-causal or prefix language model, the model attends bidirectionally over input tokens and only causally over target tokens. Given a pretrained causal decoder, previous work found that multitask finetuning in a non-causal setup performed better than causal finetuning Wang et al. (2022a); Tay et al. (2022b). However, in our experiments, non-causal finetuning did not improve over causal finetuning.
Special tokens
Instead of separating inputs and targets with a space, we experimented with special tokens. Using the end-of-sequence token as a separator or a completely new token that the model would learn during finetuning significantly worsened results. The models may need to train on more tokens, possibly even during pretraining, to learn these new special tokens Zeng et al. (2022).
Fixing prompts
PromptSource has been written with encoder-decoder models in mind, where inputs and targets are fed into different models. As a consequence, human-written prompts in PromptSource often lack separators between input and target. For our decoder models, we decided to separate them with a space. We additionally experimented with leaving them as is or rewriting a significant amount of prompts, but neither improved significantly over space separation.
BitFit
Previous work has shown bias-only finetuning Zaken et al. (2021) of large language models to be sufficient for strong downstream performance Logan et al. (2021); Hu et al. (2021); Muennighoff (2022); Liu et al. (2022); Ding et al. (2022); Muennighoff et al. (2022). We found multitask finetuning of only biases to perform 15 absolute percentage points worse on the average of held-out tasks for BLOOMZ-7.1B.
Appendix J Limitations
We highlight several limitations of our work:
Unnatural prompting format
The choice to separate inputs and targets using a space character has proven effective to multitask finetune our decoder-only models. Nonetheless, poorly formatted prompts may result in undesirable behavior. For example, given the following prompt: “Translate to English: Je t’aime”, the model may continue the input with additional French content before starting to solve the task, i.e. translating the input from French to English. This can be mitigated by improving the prompts with a trailing full stop or a newline symbol. Encoder-decoder models, such as our mT0, do not suffer from this problem, as inputs and targets are fed into different parts of the model.
Limited languages xP3
The pretraining corpus of mT0 contains more than 101 languages Xue et al. (2020), however we finetune on only 46 languages. As shown in Appendix §B, more datasets lead to better performance. Likely, extending xP3 to the full 101 languages mT0 has seen during pretraining would lead to better performance. However, we decided to use only the languages of BLOOM in order to study language generalization (§4.2). Similarly, one could likely attain better performance by enhancing xP3 with more datasets, such as via BIG-Bench Srivastava et al. (2022); Suzgun et al. (2022), or more prompts, such as via NL-Augmenter Dhole et al. (2021).
Performance
While our models show strong capabilities of performing tasks zero-shot, there remain numerous failure modes that are common in large language models Rae et al. (2021); Bommasani et al. (2021); Zhang et al. (2022); Smith et al. (2022); Ouyang et al. (2022); Chowdhery et al. (2022). In Figure 16 of Appendix §F, BLOOMZ fails to understand the moral of a fable resulting in an undesirable generation. Similarly, in Figure 15, mT0-13B is asked to provide an explanation, but answers with a question.
Learning new languages during finetuning
While we investigated generalization to languages only seen during pretraining, we did not investigate generalization to languages only seen during finetuning. Our mT0 models are finetuned on several new languages not seen in pretraining (see Figure 2). Out of those, we only evaluated on code (HumanEval), where mT0 performed at the random baseline (0.00 in Table 7). Future work may investigate language acquisition via crosslingual multitask finetuning. We point to prior work that has looked into extending BLOOM to new languages Yong and Nikoulina (2022).
Appendix K Full results
Table 7 shows all experimental results reported in this paper.
share
