AI Paper Atlas

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

AAAI 2026 · Large language models › Language, safety and interpretability · approach: Faithfulness of LLM explanations · 0 citations (OpenAlex)

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 changedWhereMeasuredEffect
add micro tokens (ablation)BEAT · AmazonBARTScore: -4.0600 -> -3.9929Improved
add micro tokens (ablation)BEAT · GoogleBLEU: 0.3027 -> 0.3866Improved
add macro tokens (ablation)BEAT · GoogleBLEU: 0.3834 -> 0.3866No clear effect
add macro tokens (ablation)BEAT · AmazonBERTScore: 0.3415 -> 0.3821 (Amazon BERT); Yelp BLEU 0.3852 -> 0.3771 hurtsMixed
add semantic alignment regularization (SAR) (ablation)BEAT · YelpBLEU: 0.2967 -> 0.3771Improved
add semantic alignment regularization (SAR) (ablation)BEAT · GoogleBLEU: 0.3922 -> 0.3866Hurt
BEAT vs strongest baselineBEAT · AmazonBLEU: 0.3971 -> 0.4195Improved
BEAT vs PLETER (best baseline BLEU)BEAT · GoogleBLEU: 0.3850 -> 0.3866No clear effect
swap LLaMA3-8B for DeepSeek-8B or Skywork-8B backbone (ablation)BEAT · AmazonBARTScore/BERTScoreHurt
swap LLaMA3-8B for LLaMA3.2-3B (ablation)BEAT · AmazonBARTScore/BERTScoreNo clear effect

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy