Research map / Large language models / Alignment and preferences
Exploring Implicit Feedback for Open Domain Conversation Generation
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 implicit feedback reward (stance, sentiment, stalemate) to objective reward (ablation) | RL-Seq2Seq · OpenSubtitles | simulated turns: 4.38 -> 4.64 | Helped |
| add implicit feedback reward (ablation) | RL-Seq2Seq · Twitter | simulated turns: 7.32 -> 7.79 | Helped |
| use implicit reward only instead of objective reward (ablation) | RL-Seq2Seq · OpenSubtitles | simulated turns: 4.38 -> 3.71 | Hurt |
| add implicit feedback reward (ablation) | RL-Seq2Seq · OpenSubtitles | distinct bigram ratio: 0.0038 -> 0.0045 | Helped |
| use proposed model vs RL-Seq2Seq in online conversation | RL-Seq2Seq · Online human eval | user satisfaction: 1.974 -> 2.304 | Helped |
| use proposed model vs RL-Seq2Seq in single-turn judgement | RL-Seq2Seq · Single-turn human judgement | win rate: 0.153 win vs 0.12 lose, 0.726 tie | Mixed |
Closest papers by meaning
- ULTRAFEEDBACK: Boosting Language Models with Scaled AI Feedback (ICML 2024 · similarity 0.86)
- Towards Better Robot Learners: Leveraging Implicit and Explicit Human Feedback Together in Human Robot Interactions (AAAI 2025 · similarity 0.85)
- Decoupling Conversational Dynamics in Full-Duplex Spoken Models through Reinforcement Learning (NeurIPS 2026 · similarity 0.85)
- MA-RLHF: Reinforcement Learning from Human Feedback with Macro Actions (ICLR 2025 · similarity 0.85)
- Enabling Lanuguage Models to Implicitly Learn Self-Improvement (ICLR 2024 · similarity 0.85)
- Aligning Language Models Using Follow-up Likelihood as Reward Signal (AAAI 2025 · similarity 0.85)
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.