FAST THINKING FOR LARGE LANGUAGE MODELS

Haoyu Zheng, Zhuonan Wang, Yuqian Yuan, Wenqiao Zhang, Zheqi Lv, Hongyang He, Juncheng Li, Siliang Tang, Yueting Zhuang

Keywords: Large Language Models, Efficient Reasoning, Adaptive Routing

TL;DR: LC-FT learns a latent codebook of strategy priors from concise CoT, then uses a few continuous hint vectors to enable single-pass “fast thinking.”

Abstract:

Reasoning-oriented Large Language Models (LLMs) often rely on generating explicit tokens step by step, and their effectiveness typically hinges on large-scale supervised fine-tuning or reinforcement learning.
While Chain-of-Thought (CoT) techniques substantially enhance performance on complex reasoning tasks, they remain inefficient, requiring long reasoning traces that increase latency and token usage.
In this work, we introduce \emph{Latent Codebooks for Fast Thinking}, a framework that uses concise CoT sketches only during training to learn a codebook of discrete strategy priors.
At inference, the model conditions on a handful of continuous thinking vectors distilled from the codebook in a single pass, enabling strategy-level guidance without producing explicit reasoning tokens.
To complement this design, we propose GainRouter, a lightweight routing mechanism that adaptively switches between fast codebook-guided inference and slow explicit reasoning, thereby suppressing overthinking and reducing unnecessary token generation.
Experiments across multiple reasoning benchmarks show that our approach achieves competitive or superior accuracy while substantially lowering inference cost, offering a practical path toward efficient and controllable reasoning in large language models.

Primary Area: foundation or frontier models, including LLMs

Submission Number: 10249