TFRank: Think-Free Reasoning Enables Practical Pointwise LLM Ranking
Yongqi Fan, Xiaoyang Chen, Dezhi Ye, Jie Liu, Haijin Liang, Jin Ma, Ben He, Yingfei Sun, Tong Ruan

TL;DR
TFRank is an efficient pointwise reasoning ranker based on small-scale LLMs that integrates reasoning data and multi-task training to achieve high performance with reduced computational costs, enabling practical real-world applications.
Contribution
The paper introduces TFRank, a novel reasoning ranker that combines explicit reasoning with a think-free inference mode using small LLMs, reducing costs while maintaining high accuracy.
Findings
TFRank matches larger models' performance on BRIGHT benchmark.
It demonstrates strong competitiveness on the BEIR benchmark.
Achieves a balance between efficiency and ranking accuracy.
Abstract
Reasoning-intensive ranking models built on Large Language Models (LLMs) have made notable progress. However, existing approaches often rely on large-scale LLMs and explicit Chain-of-Thought (CoT) reasoning, resulting in high computational cost and latency that limit real-world use. To address this, we propose \textbf{TFRank}, an efficient pointwise reasoning ranker based on small-scale LLMs. To improve ranking performance, TFRank effectively integrates CoT data, fine-grained score supervision, and multi-task training. Furthermore, it achieves an efficient ``\textbf{T}hink-\textbf{F}ree" reasoning capability by employing a ``think-mode switch'' and pointwise format constraints. Specifically, this allows the model to leverage explicit reasoning during training while delivering precise relevance scores for complex queries at inference without generating any reasoning chains. Experiments…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsSemantic Web and Ontologies
