FLEx: Language Modeling with Few-shot Language Explanations

TL;DR

FLEx is a method that enhances language model accuracy by using a small set of explanatory examples, selected through clustering, to guide the model and reduce errors without changing its weights.

Contribution

FLEx introduces a novel approach for improving language model performance with few-shot explanations, leveraging embedding-based clustering and error correction summaries.

Findings

FLEx outperforms chain-of-thought prompting on multiple datasets.

FLEx reduces up to 83% of errors compared to CoT.

The method does not require modifying model weights.

Abstract

Language models have become effective at a wide range of tasks, from math problem solving to open-domain question answering. However, they still make mistakes, and these mistakes are often repeated across related queries. Natural language explanations can help correct these errors, but collecting them at scale may be infeasible, particularly in domains where expert annotators are required. To address this issue, we introduce FLEx ($\textbf{F}$ew-shot $\textbf{L}$anguage $\textbf{Ex}$planations), a method for improving model behavior using a small number of explanatory examples. FLEx selects representative model errors using embedding-based clustering, verifies that the associated explanations correct those errors, and summarizes them into a prompt prefix that is prepended at inference-time. This summary guides the model to avoid similar errors on new inputs, without modifying model weights. We evaluate FLEx on CounterBench, GSM8K, and ReasonIF. We find that FLEx consistently outperforms chain-of-thought (CoT) prompting across all three datasets and reduces up to 83\% of CoT's remaining errors.