Research map / Large language models / Retrieval and knowledge
Knowledge-Driven Distractor Generation for Cloze-Style Multiple Choice Questions
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 changed | Where | Measured | Effect |
|---|---|---|---|
| Add the list-wise distractor selector to Probase CSG (ablation) | Probase CSG · Test set | F1@3: 5.88 -> 9.19 | Helped |
| Add the list-wise distractor selector to WordNet CSG (ablation) | WordNet CSG · Test set | F1@3: 3.14 -> 7.71 | Helped |
| Use Probase rather than WordNet as the candidate-set knowledge base (ablation) | CSG + DS (list-wise) · Test set | F1@3: 7.71 -> 9.19 | Helped |
| Remove CSG and use an unstructured candidate set (ablation) | Distractor selector · Test set | F1@3: 9.19 -> 5.94 | Hurt |
| Use a point-wise rather than list-wise ranker (ablation) | Probase CSG + DS · Test set | F1@3: 9.19 -> 7.91 | Hurt |
| Use a pair-wise rather than list-wise ranker (ablation) | Probase CSG + DS · Test set | F1@3: 9.19 -> 9.42 | Mixed |
| Use a point-wise rather than list-wise ranker (ablation) | WordNet CSG + DS · Test set | F1@3: 7.71 -> 7.26 | Hurt |
| Use a pair-wise rather than list-wise ranker (ablation) | WordNet CSG + DS · Test set | F1@3: 7.71 -> 7.11 | Hurt |
| Use DS rather than LR+RF for candidate ranking | Probase CSG · Test set | F1@3: 7.48 -> 9.19 | Helped |
| Use DS rather than LR+LM for candidate ranking | Probase CSG · Test set | F1@3: 7.62 -> 9.19 | Helped |
| Use DS rather than EmbSim+CF for candidate ranking | Probase CSG · Test set | F1@3: 7.01 -> 9.19 | Helped |
| Use edit distance to rank candidates | Probase CSG · Test set | F1@3: 5.88 -> 0.82 | Hurt |
Closest papers by meaning
- Automated Generation of Challenging Multiple-Choice Questions for Vision Language Model Evaluation (CVPR 2025 · similarity 0.83)
- Adaptive Distraction: Probing LLM Contextual Robustness with Automated Tree Search (NeurIPS 2025 · similarity 0.83)
- Rethinking Generative Large Language Model Evaluation for Semantic Comprehension (ICML 2024 · similarity 0.83)
- Answer, Assemble, Ace: Understanding How LMs Answer Multiple Choice Questions (ICLR 2025 · similarity 0.83)
- Reverse Multi-Choice Dialogue Commonsense Inference with Graph-of-Thought (AAAI 2024 · similarity 0.82)
- DSAS: A Universal Plug-and-Play Framework for Attention Optimization in Multi-Document Question Answering (NeurIPS 2025 · similarity 0.82)
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.