AI Paper Atlas

Research map / Large language models / Retrieval and knowledge

Counterfactual Question Generation Uncovering Learner Contradictions

Bo Zhang, Hao Yu, Wenjie Dong, Yvhang Yang, Dezhuang Miao, Fengyi Song, Yanhui Gu, Xiaoming Zhang, Junsheng Zhou

AAAI 2026 · Large language models › Retrieval and knowledge · approach: Explanatory and multiple-choice QA benchmarks · 0 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
add KGMap schema memory to GapProbe (ablation)GapProbe · CWQ-CFQG testCognitive Depth (DeepSeek-V3): 0.748 -> 0.949Improved
add KGMap schema memory to GapProbe (ablation)GapProbe · CWQ-CFQG testGood CFQ rate (%) (DeepSeek-V3): 49.5 -> 62.1Improved
add KGMap schema memory to GapProbe (ablation)GapProbe · CWQ-CFQG testGood CFQ rate (%) (GPT-4): 42.3 -> 47.6Improved
add KGMap schema memory to GapProbe (ablation)GapProbe · CWQ-CFQG testReflection Trigger (DeepSeek-V3): 0.738 -> 0.672Hurt
add KGMap schema memory to GapProbe (ablation)GapProbe · CWQ-CFQG testKnowledge Focus: 0.937 -> 0.963 (GPT-4); 0.972 -> 0.960 (DS-v3)No clear effect
use ProConB iterative cycle instead of MFC-style single passCFQ generation · CWQ-CFQG testReflection Trigger (GPT-4): 0.377 -> 0.505 (w/o KGMap)Improved
add confusability filtering (+P) (ablation)GapProbe · CWQ-CFQG testRT and Good rate; KFMixed
add multi-hop SPARQL generation (ablation)GapProbe · CWQ-CFQG testCognitive DepthImproved
use GapProbe-QA instead of MFCGPT-4 KBQA · KBQA average of four benchmarksEM accuracy (%): 60.9 -> 66.9Improved
use GapProbe-QA instead of ToGGPT-4 KBQA · CWQ-QQAEM accuracy (%): 58.1 -> 55.5Hurt
use GapProbe-QA instead of ToGDeepSeek-V3 KBQA · CWQEM accuracy (%): 56.3 -> 58.1 (ToG vs GapProbe-QA); MFC 60.0Mixed
add CFQ feedback to student LLM reflectionLlama-3-70B · LLM-as-student simulationcorrection rate (%): 8.70 -> 33.3Improved

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