Language Models are Hidden Reasoners: Unlocking Latent Reasoning
  Capabilities via Self-Rewarding

Haolin Chen; Yihao Feng; Zuxin Liu; Weiran Yao; Akshara Prabhakar; Shelby Heinecke; Ricky Ho; Phil Mui; Silvio Savarese; Caiming Xiong; Huan Wang

Paper · arXiv 2411.04282

Language Models are Hidden Reasoners: Unlocking Latent Reasoning Capabilities via Self-Rewarding

Haolin Chen, Yihao Feng, Zuxin Liu, Weiran Yao, Akshara Prabhakar, Shelby Heinecke, Ricky Ho, Phil Mui, Silvio Savarese, Caiming Xiong, Huan Wang

37 upvotesNovember 6, 2024arXiv 预印本

AI 摘要

LaTent Reasoning Optimization (LaTRO) enhances LLMs' reasoning capabilities through variational optimization of latent distributions, improving zero-shot accuracy on complex reasoning tasks.

Large language models (LLMs)Chain-of-Thought (CoT)LaTent Reasoning Optimization (LaTRO)latent distributionvariational approachesGSM8KARC-Challengezero-shot accuracyPhi-3.5-miniMistral-7BLlama-3.1-8Bpre-trained LLMslatent reasoning capabilitiesself-improvement

Abstract

Large language models (LLMs) have shown impressive capabilities, but still struggle with complex reasoning tasks requiring multiple steps. While prompt-based methods like Chain-of-Thought (CoT) can improve LLM reasoning at inference time, optimizing reasoning capabilities during training remains challenging. We introduce LaTent Reasoning Optimization (LaTRO), a principled framework that formulates reasoning as sampling from a latent distribution and optimizes it via variational approaches. LaTRO enables LLMs to concurrently improve both their reasoning process and ability to evaluate reasoning quality, without requiring external feedback or reward models. We validate LaTRO through experiments on GSM8K and ARC-Challenge datasets using multiple model architectures. On GSM8K, LaTRO improves zero-shot accuracy by an average of 12.5% over base models and 9.6% over supervised fine-tuning across Phi-3.5-mini, Mistral-7B, and Llama-3.1-8B. Our findings suggest that pre-trained LLMs possess latent reasoning capabilities that can be unlocked and enhanced through our proposed optimization approach in a self-improvement manner. The code of LaTRO is available at https://github.com/SalesforceAIResearch/LaTRO.