Tailoring Self-rationalizers With Multi-reward Distillation

Ramnath Sahana, Joshi Brihi, Hallinan Skyler, Lu Ximing, Li Liunian Harold, Chan Aaron, Hessel Jack, Choi Yejin, Ren Xiang. The Twelfth International Conference on Learning Representations 2023

[Paper]    
Applications Distillation Efficiency And Optimization Fine Tuning GPT Model Architecture Pretraining Methods Reinforcement Learning Training Techniques

Large language models (LMs) are capable of generating free-text rationales to aid question answering. However, prior work 1) suggests that useful self-rationalization is emergent only at significant scales (e.g., 175B parameter GPT-3); and 2) focuses largely on downstream performance, ignoring the semantics of the rationales themselves, e.g., are they faithful, true, and helpful for humans? In this work, we enable small-scale LMs (approx. 200x smaller than GPT-3) to generate rationales that not only improve downstream task performance, but are also more plausible, consistent, and diverse, assessed both by automatic and human evaluation. Our method, MaRio (Multi-rewArd RatIOnalization), is a multi-reward conditioned self-rationalization algorithm that optimizes multiple distinct properties like plausibility, diversity and consistency. Results on five difficult question-answering datasets StrategyQA, QuaRel, OpenBookQA, NumerSense and QASC show that not only does MaRio improve task accuracy, but it also improves the self-rationalization quality of small LMs across the aforementioned axes better than a supervised fine-tuning (SFT) baseline. Extensive human evaluations confirm that MaRio rationales are preferred vs. SFT rationales, as well as qualitative improvements in plausibility and consistency.

Similar Work