Are Language Models Efficient Reasoners? A Perspective from Logic Programming


Loading...

Date

2025-10-29

Publication Type

Conference Paper

ETH Bibliography

yes

Citations

Altmetric

Data

Abstract

Modern language models (LMs) exhibit strong deductive reasoning capabilities, yet standard evaluations emphasize correctness while overlooking a key aspect of human-like reasoning: efficiency. In real-world reasoning scenarios, much of the available information is irrelevant, and effective deductive inference requires identifying and ignoring such distractions. We propose a framework for assessing LM reasoning efficiency through the lens of logic programming, introducing a simple method to align proofs written in natural language -- as generated by an LM -- with shortest proofs found by executing the logic program. Efficiency is quantified by measuring how well a model avoids unnecessary inference. Empirically, we construct a dataset of math word problems injected with various number of irrelevant axioms that vary in semantic overlap with the goal theorem. We find that current LMs show marked accuracy declines under such conditions -- even with minimal, domain-consistent distractions -- and the proofs they generate frequently exhibit detours through irrelevant inferences.

Publication status

published

Editor

Book title

Journal / series

Volume

Pages / Article No.

22305

Publisher

OpenReview

Event

39th Annual Conference on Neural Information Processing Systems (NeurIPS 2025)

Edition / version

Methods

Software

Geographic location

Date collected

Date created

Subject

Language model; Logic programming; Proof; Arithmetic reasoning; Math word problem; Search; Reasoning; Efficiency

Organisational unit

09682 - Cotterell, Ryan / Cotterell, Ryan

Notes

Postersession on December 5, 2025

Funding

Related publications and datasets