Failures Pave the Way: Enhancing Large Language Models through Tuning-free Rule Accumulation
This addresses a specific issue for LLM users by enhancing model accuracy without retraining, though it appears incremental as it builds on existing prompt-based methods.
The paper tackles the problem of large language models (LLMs) repeating mistakes due to an inability to capture sample relationships, proposing a tuning-free rule accumulation framework that learns from errors to improve performance, with experiments showing large-margin improvements over baselines.
Large Language Models (LLMs) have showcased impressive performance. However, due to their inability to capture relationships among samples, these frozen LLMs inevitably keep repeating similar mistakes. In this work, we propose our Tuning-free Rule Accumulation (TRAN) framework, which guides LLMs in improving their performance by learning from previous mistakes. Considering data arrives sequentially, LLMs gradually accumulate rules from incorrect cases, forming a rule collection. These rules are then utilized by the LLMs to avoid making similar mistakes when processing subsequent inputs. Moreover, the rules remain independent of the primary prompts, seamlessly complementing prompt design strategies. Experimentally, we show that TRAN improves over recent baselines by a large margin.