Research map / Large language models / Retrieval and knowledge
Response Generation by Context-Aware Prototype Editing
Yu Wu, Furu Wei, Shaohan Huang, Yunli Wang, Zhoujun Li, Ming Zhou
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 changed | Where | Measured | Effect |
|---|---|---|---|
| Use a prototype response and context-aware editing instead of generating from scratch | Response generation · Douban test | Not appear: 0.745 -> 0.860 | Improved |
| Use prototype-then-edit rather than a traditional generative model | Response generation · Douban test | Distinct-2: 0.178 -> 0.280 | Improved |
| Edit retrieved candidates and rerank them | Retrieval response generation · Douban test | Average: 0.380 -> 0.386 | Improved |
| Edit one top-ranked retrieval response before reranking | Retrieval response generation · Douban test | Average: 0.380 -> 0.367 | Hurt |
| Add context insertion words to the edit vector (ablation) | Context-aware neural editor · Douban test | Average: 34.2 -> 38.6 | Improved |
| Add context deletion words to the edit vector (ablation) | Context-aware neural editor · Douban test | Average: 34.0 -> 38.6 | Improved |
| Add context-aware insertion and deletion information rather than standard Seq2Seq editing (ablation) | Context-aware neural editor · Douban test | Average: 32.8 -> 38.6 | Improved |
| Add context insertion words to the edit vector (ablation) | Context-aware neural editor · Douban test | Originality: 85.1 -> 86.0 | No clear effect |
| Add context deletion words to the edit vector (ablation) | Context-aware neural editor · Douban test | Originality: 86.3 -> 86.0 | No clear effect |
| Add context-aware insertion and deletion information (ablation) | Context-aware neural editor · Douban test | Dis-2: 0.282 -> 0.280 | No clear effect |
Closest papers by meaning
- Retrieval is Accurate Generation (ICLR 2024 · similarity 0.86)
- Provence: efficient and robust context pruning for retrieval-augmented generation (ICLR 2025 · similarity 0.85)
- Context Steering: Controllable Personalization at Inference Time (ICLR 2025 · similarity 0.85)
- RevisEval: Improving LLM-as-a-Judge via Response-Adapted References (ICLR 2025 · similarity 0.84)
- Influence Guided Context Selection for Effective Retrieval-Augmented Generation (NeurIPS 2025 · similarity 0.84)
- In-Context Editing: Learning Knowledge from Self-Induced Distributions (ICLR 2025 · similarity 0.84)
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.
