Forgetful causal masking makes causal language models better zero-shot learners

Hao Liu, Xinyang Geng, Lisa Lee, Igor Mordatch, Sergey Levine, Sharan Narang, Pieter Abbeel

Keywords: Language modeling, casual language model, few shot language models

Abstract: Large language models (LLM) trained using the next-token-prediction objective, such as GPT3 and PaLM, have revolutionized natural language processing in recent years by showing impressive zero-shot and few-shot capabilities across a wide range of tasks. In this work, we propose a simple technique that significantly boosts the performance of LLMs without adding computational cost. Our key observation is that, by performing the next token prediction task with randomly selected past tokens masked out, we can improve the quality of the learned representations for downstream language understanding tasks. We hypothesize that randomly masking past tokens prevents over-attending to recent tokens and encourages attention to tokens in the distant past. By randomly masking input tokens in the PaLM model, we show that we can significantly improve PaLM's zero-shot performance on the SuperGLUE benchmark from 55.7 to 59.2. Experimental results show that FCM also improves PaLM's zero- and few-shot performance on a diverse suite of tasks, including commonsense reasoning, natural language inference and cloze completion. Moreover, we show that our technique also helps representation learning, significantly improving PaLM's finetuning results on SuperGLUE.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics

Submission Guidelines: Yes

Please Choose The Closest Area That Your Submission Falls Into: Deep Learning and representational learning

TL;DR: A simple masking strategy for casual language models that significantly improves few shot and finetuning performance.