Research map / Large language models / Language, safety and interpretability
Do Response Selection Models Really Know What’s Next? Utterance Manipulation Strategies for Multi-turn Response Selection
Taesun Whang, Dongyub Lee, Dongsuk Oh, Chanhee Lee, Kijong Han, Dong-hun Lee, Saebyeok Lee
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 |
|---|---|---|---|
| Add insertion, deletion, and search tasks to domain-post-trained BERT | BERT+ · Ubuntu | R_{10}@1: 0.862 -> 0.875 | Improved |
| Add insertion, deletion, and search tasks to domain-post-trained BERT | BERT+ · Douban | MAP: 0.609 -> 0.625 | Improved |
| Add insertion, deletion, and search tasks to domain-post-trained BERT | BERT+ · Douban | MRR: 0.645 -> 0.664 | Improved |
| Add insertion, deletion, and search tasks to domain-post-trained BERT | BERT+ · E-Commerce | R_{10}@1: 0.725 -> 0.762 | Improved |
| Add insertion, deletion, and search tasks | BERT · Kakao Web | P@1: 0.555 -> 0.606 | Improved |
| Add insertion, deletion, and search tasks | BERT · Kakao Clean | P@1: 0.648 -> 0.716 | Improved |
| Add deletion-only auxiliary training (ablation) | ELECTRA response selection · Ubuntu | R_{10}@1: 0.826 -> 0.848 | Improved |
| Add insertion-only auxiliary training (ablation) | ELECTRA response selection · Ubuntu | R_{10}@1: 0.826 -> 0.836 | Improved |
| Add search-only auxiliary training (ablation) | ELECTRA response selection · Ubuntu | R_{10}@1: 0.826 -> 0.834 | Improved |
| Add insertion and deletion, without search (ablation) | ELECTRA response selection · Ubuntu | R_{10}@1: 0.826 -> 0.853 | Improved |
| Add insertion and search, without deletion (ablation) | ELECTRA response selection · Ubuntu | R_{10}@1: 0.826 -> 0.841 | Improved |
| Add deletion and search, without insertion (ablation) | ELECTRA response selection · Ubuntu | R_{10}@1: 0.826 -> 0.852 | Improved |
Closest papers by meaning
- An Efficient Self-Learning Framework For Interactive Spoken Dialog Systems (ICML 2024 · similarity 0.86)
- Learning to Clarify: Multi-turn Conversations with Action-Based Contrastive Self-Training (ICLR 2025 · similarity 0.86)
- Aligning Spoken Dialogue Models from User Interactions (ICML 2025 · similarity 0.85)
- A Comprehensive Analysis of the Effectiveness of Large Language Models as Automatic Dialogue Evaluators (AAAI 2024 · similarity 0.85)
- Paralinguistics-Aware Speech-Empowered Large Language Models for Natural Conversation (NeurIPS 2024 · similarity 0.84)
- A Multi-Agent Conversational Bandit Approach to Online Evaluation and Selection of User-Aligned LLM Responses (AAAI 2026 · 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.
