Self-AMPLIFY: Improving Small Language Models with Self Post Hoc Explanations
This addresses the challenge of enhancing SLM performance without human annotation or proxy models, though it is incremental in applying existing explanation techniques to a new context.
The paper tackles the problem of improving small language models (SLMs) by automatically generating rationales using post hoc explanation methods, achieving strong accuracy improvements on five reasoning datasets.
Incorporating natural language rationales in the prompt and In-Context Learning (ICL) have led to a significant improvement of Large Language Models (LLMs) performance. However, generating high-quality rationales require human-annotation or the use of auxiliary proxy models. In this work, we propose Self-AMPLIFY to automatically generate rationales from post hoc explanation methods applied to Small Language Models (SLMs) to improve their own performance. Self-AMPLIFY is a 3-step method that targets samples, generates rationales and builds a final prompt to leverage ICL. Self-AMPLIFY performance is evaluated on four SLMs and five datasets requiring strong reasoning abilities. Self-AMPLIFY achieves good results against competitors, leading to strong accuracy improvement. Self-AMPLIFY is the first method to apply post hoc explanation methods to autoregressive language models to generate rationales to improve their own performance in a fully automated manner.