✨Extracting Interaction-Aware Monosemantic Concepts in Recommender Systems

📝 Summary:
A Sparse Autoencoder extracts interaction-aware monosemantic concepts from recommender embeddings. Its prediction-aware training aligns these with model predictions, enabling controllable personalization and interpretability.

🔹 Publication Date: Published on Nov 22

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2511.18024
• PDF: https://arxiv.org/pdf/2511.18024

==================================

For more data science resources:
✓ https://t.me/DataScienceT

#RecommenderSystems #DeepLearning #AI #Interpretability #Personalization

✨Fantastic Reasoning Behaviors and Where to Find Them: Unsupervised Discovery of the Reasoning Process

📝 Summary:
This paper introduces RISE, an unsupervised framework using sparse auto-encoders to discover and control LLM reasoning behaviors. It identifies interpretable reasoning vectors like reflection and backtracking, enabling targeted interventions and discovery of novel behaviors without retraining.

🔹 Publication Date: Published on Dec 30, 2025

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2512.23988
• PDF: https://arxiv.org/pdf/2512.23988

==================================

For more data science resources:
✓ https://t.me/DataScienceT

#LLM #AI #MachineLearning #AIReasoning #Interpretability