Best AI papers explained

518 Episodes

1. Learning to summarize user information for personalized reinforcement learning from human feedback

   Published: 4/10/2025
2. Distributional Preference Learning: Understanding and Accounting for Hidden Context in RLHF

   Published: 3/10/2025
3. LIMI: Less is More for Agency

   Published: 1/10/2025
4. LoRA Without Regret

   Published: 1/10/2025
5. Actor-Critic without Actor: Critic-Guided Denoising for RL

   Published: 29/09/2025
6. DELTA-Code: How Does RL Unlock and Transfer New Programming Algorithms in LLMs?

   Published: 29/09/2025
7. Linear Transformers Implicitly Discover Unified Numerical Algorithms

   Published: 29/09/2025
8. Regularizing Extrapolation in Causal Inference

   Published: 27/09/2025
9. DoubleGen - Debiased Generative Modeling of Counterfactuals

   Published: 27/09/2025
10. What Characterizes Effective Reasoning? Revisiting Length, Review, and Structure of CoT

    Published: 27/09/2025
11. Compute as Teacher: Turning Inference Compute Into Reference-Free Supervision

    Published: 27/09/2025
12. Learning without training: The implicit dynamics of in-context learning

    Published: 24/09/2025
13. Does Reinforcement Learning Really Incentivize Reasoning Capacity in LLMs Beyond the Base Model

    Published: 24/09/2025
14. Open Problems in Mechanistic Interpretability

    Published: 21/09/2025
15. Maestro: Joint Graph & Config Optimization for Reliable AI Agents

    Published: 21/09/2025
16. Thought Anchors: Which LLM Reasoning Steps Matter?

    Published: 21/09/2025
17. Sample Complexity and Representation Ability of Test-time Scaling Paradigms

    Published: 9/09/2025
18. RL's Razor: Why Online RL Forgets Less

    Published: 7/09/2025
19. Why Language Models Hallucinate

    Published: 6/09/2025
20. ALFA: Aligning LLMs to Ask Good Questions A Case Study in Clinical Reasoning

    Published: 6/09/2025