Best AI papers explained

522 Episoade

1. Language Bottleneck Models: A Framework for Interpretable Knowledge Tracing and Beyond

   Publicat: 03.07.2025
2. Learning to Explore: An In-Context Learning Approach for Pure Exploration

   Publicat: 03.07.2025
3. Human-AI Matching: The Limits of Algorithmic Search

   Publicat: 25.06.2025
4. Uncertainty Quantification Needs Reassessment for Large-language Model Agents

   Publicat: 25.06.2025
5. Bayesian Meta-Reasoning for Robust LLM Generalization

   Publicat: 25.06.2025
6. General Intelligence Requires Reward-based Pretraining

   Publicat: 25.06.2025
7. Deep Learning is Not So Mysterious or Different

   Publicat: 25.06.2025
8. AI Agents Need Authenticated Delegation

   Publicat: 25.06.2025
9. Probabilistic Modelling is Sufficient for Causal Inference

   Publicat: 25.06.2025
10. Not All Explanations for Deep Learning Phenomena Are Equally Valuable

    Publicat: 25.06.2025
11. e3: Learning to Explore Enables Extrapolation of Test-Time Compute for LLMs

    Publicat: 17.06.2025
12. Extrapolation by Association: Length Generalization Transfer in Transformers

    Publicat: 17.06.2025
13. Uncovering Causal Hierarchies in Language Model Capabilities

    Publicat: 17.06.2025
14. Generalization or Hallucination? Understanding Out-of-Context Reasoning in Transformers

    Publicat: 17.06.2025
15. Improving Treatment Effect Estimation with LLM-Based Data Augmentation

    Publicat: 17.06.2025
16. LLM Numerical Prediction Without Auto-Regression

    Publicat: 17.06.2025
17. Self-Adapting Language Models

    Publicat: 17.06.2025
18. Why in-context learning models are good few-shot learners?

    Publicat: 17.06.2025
19. Take Caution in Using LLMs as Human Surrogates: Scylla Ex Machina∗

    Publicat: 14.06.2025
20. The Logic of Machines: The AI Reasoning Debate

    Publicat: 12.06.2025