AI Paper Atlas

Research map / Large language models / Agents, code and math

Socrates or Smartypants: Testing Logic Reasoning Capabilities of Large Language Models with Logic Programming-Based Test Oracles

Zihao Xu, Junchen Ding, Yiling Lou, Kun Zhang, Dong Gong, Yuekang Li

AAAI 2026 · Large language models › Agents, code and math · approach: Mathematical and structural reasoning benchmarks · 4 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
generate fallacies with Prolog rules and LLM-filled facts instead of direct LLM generationSmartyPat vs FallacyGen-Direct · Sentence quality, GPT-4o judgeaverage quality score: +38.12% on averageImproved
constrain generation with Prolog predicates and validation instead of unconstrained LLM-written Prolog (ablation)SmartyPat vs FallacyGen-Prolog · Fallacy of Composition, score-3 countnumber of score-3 sentences: 25 -> 60Improved
group facts per fallacy instance instead of per predicate in few-shot prompts (ablation)Prolog knowledge generation prompt · LLM fact generationfact coherence (qualitative)Improved
add inline comments clarifying each predicate in few-shot facts (ablation)Prolog knowledge generation prompt · LLM fact generationLLM understanding (qualitative)Improved
use reasoning models instead of non-reasoning models for fallacy detectionnine LLMs · SmartyPat-Bench detectionF1: DeepSeek V3 0.934, Grok-2 0.921 vs GPT-o3-mini 0.891Hurt
use reasoning models for fallacy categorizationnine LLMs · SmartyPat-Bench categorizationrank-weighted label score: GPT-o3-mini -8.41 vs GPT-4o -126.46Improved
evaluate on synthetic SmartyPat-Bench-Augmented instead of real-world SmartyPat-BenchLLM fallacy categorization · categorization scorerank-weighted label score: Claude 3.7 improves by over 70%Improved
evaluate on synthetic SmartyPat-Bench-Augmented instead of real-world SmartyPat-BenchLLM fallacy detection · detection F1FPR/FNR/F1No clear effect
move from surface-level to context-dependent fallacy typesLLM fallacy detection · SmartyPat-Bench-Augmenteddetection accuracyHurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy