Research map / Large language models / Language, safety and interpretability
Behavior Tokens Speak Louder: Disentangled Explainable Recommendation with Behavior Vocabulary
Xinshun Feng, Mingzhe Liu, Yi Qiao, Tongyu Zhu, Leilei Sun, Shuai Wang
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 micro tokens (ablation) | BEAT · Amazon | BARTScore: -4.0600 -> -3.9929 | Improved |
| add micro tokens (ablation) | BEAT · Google | BLEU: 0.3027 -> 0.3866 | Improved |
| add macro tokens (ablation) | BEAT · Google | BLEU: 0.3834 -> 0.3866 | No clear effect |
| add macro tokens (ablation) | BEAT · Amazon | BERTScore: 0.3415 -> 0.3821 (Amazon BERT); Yelp BLEU 0.3852 -> 0.3771 hurts | Mixed |
| add semantic alignment regularization (SAR) (ablation) | BEAT · Yelp | BLEU: 0.2967 -> 0.3771 | Improved |
| add semantic alignment regularization (SAR) (ablation) | BEAT · Google | BLEU: 0.3922 -> 0.3866 | Hurt |
| BEAT vs strongest baseline | BEAT · Amazon | BLEU: 0.3971 -> 0.4195 | Improved |
| BEAT vs PLETER (best baseline BLEU) | BEAT · Google | BLEU: 0.3850 -> 0.3866 | No clear effect |
| swap LLaMA3-8B for DeepSeek-8B or Skywork-8B backbone (ablation) | BEAT · Amazon | BARTScore/BERTScore | Hurt |
| swap LLaMA3-8B for LLaMA3.2-3B (ablation) | BEAT · Amazon | BARTScore/BERTScore | No clear effect |
Closest papers by meaning
- Semantic Convergence: Harmonizing Recommender Systems via Two-Stage Alignment and Behavioral Semantic Tokenization (AAAI 2025 · similarity 0.89)
- Language Representations Can be What Recommenders Need: Findings and Potentials (ICLR 2025 · similarity 0.88)
- FACE: A General Framework for Mapping Collaborative Filtering Embeddings into LLM Tokens (NeurIPS 2025 · similarity 0.88)
- Exploring Large Language Model for Graph Data Understanding in Online Job Recommendations (AAAI 2024 · similarity 0.87)
- Bridging the User-side Knowledge Gap in Knowledge-aware Recommendations with Large Language Models (AAAI 2025 · similarity 0.86)
- ActionPiece: Contextually Tokenizing Action Sequences for Generative Recommendation (ICML 2025 · similarity 0.86)
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.
