What question did this study set out to answer?

This work critically examines the effectiveness of finetuning weaker language models using outputs from stronger proprietary models.

May 25, 202351 citationsOpen Access

The False Promise of Imitating Proprietary LLMs

AGArnav Gudibande EWEric WallaceUniversity of Alabama at Birmingham Hospital CSCharlie SnellUniversity of California, Berkeley

Key Points

This work critically examines the effectiveness of finetuning weaker language models using outputs from stronger proprietary models.
Finetuned multiple language models imitating ChatGPT with varying sizes (1.5B-13B) and imitation data amounts (0.3M-150M tokens).
Evaluated model outputs using crowd raters and standard NLP benchmarks to assess performance.
Conducted targeted automatic evaluations to analyze discrepancies in model performance.
Imitation models rated competitively by crowd workers but failed to replicate ChatGPT's factuality in automated tests.
Performance gap remains significant for tasks outside the imitation data support.
Concludes imitation methods are insufficient for bridging capabilities gap; improving base models is essential.

Abstract

An emerging method to cheaply improve a weaker language model is to finetune it on outputs from a stronger model, such as a proprietary system like ChatGPT (e.g., Alpaca, Self-Instruct, and others). This approach looks to cheaply imitate the proprietary model's capabilities using a weaker open-source model. In this work, we critically analyze this approach. We first finetune a series of LMs that imitate ChatGPT using varying base model sizes (1.5B--13B), data sources, and imitation data amounts (0.3M--150M tokens). We then evaluate the models using crowd raters and canonical NLP benchmarks. Initially, we were surprised by the output quality of our imitation models -- they appear far better at following instructions, and crowd workers rate their outputs as competitive with ChatGPT. However, when conducting more targeted automatic evaluations, we find that imitation models close little to none of the gap from the base LM to ChatGPT on tasks that are not heavily supported in the imitation data. We show that these performance discrepancies may slip past human raters because imitation models are adept at mimicking ChatGPT's style but not its factuality. Overall, we conclude that model imitation is a false promise: there exists a substantial capabilities gap between open and closed LMs that, with current methods, can only be bridged using an unwieldy amount of imitation data or by using more capable base LMs. In turn, we argue that the highest leverage action for improving open-source models is to tackle the difficult challenge of developing better base LMs, rather than taking the shortcut of imitating proprietary systems.

AIに質問

Bookmark

View Full Paper