Best AI papers explained

528 Episoade

1. Stratified Prediction-Powered Inference for Hybrid Language Model Evaluation

   Publicat: 09.05.2025
2. Accelerating Unbiased LLM Evaluation via Synthetic Feedback

   Publicat: 09.05.2025
3. Prediction-Powered Statistical Inference Framework

   Publicat: 09.05.2025
4. Optimizing Chain-of-Thought Reasoners via Gradient Variance Minimization in Rejection Sampling and RL

   Publicat: 09.05.2025
5. RM-R1: Reward Modeling as Reasoning

   Publicat: 09.05.2025
6. Reexamining the Aleatoric and Epistemic Uncertainty Dichotomy

   Publicat: 08.05.2025
7. Decoding Claude Code: Terminal Agent for Developers

   Publicat: 07.05.2025
8. Emergent Strategic AI Equilibrium from Pre-trained Reasoning

   Publicat: 07.05.2025
9. Benefiting from Proprietary Data with Siloed Training

   Publicat: 06.05.2025
10. Advantage Alignment Algorithms

    Publicat: 06.05.2025
11. Asymptotic Safety Guarantees Based On Scalable Oversight

    Publicat: 06.05.2025
12. What Makes a Reward Model a Good Teacher? An Optimization Perspective

    Publicat: 06.05.2025
13. Towards Guaranteed Safe AI: A Framework for Ensuring Robust and Reliable AI Systems

    Publicat: 06.05.2025
14. Identifiable Steering via Sparse Autoencoding of Multi-Concept Shifts

    Publicat: 06.05.2025
15. You Are What You Eat - AI Alignment Requires Understanding How Data Shapes Structure and Generalisation

    Publicat: 06.05.2025
16. Interplay of LLMs in Information Retrieval Evaluation

    Publicat: 03.05.2025
17. Trade-Offs Between Tasks Induced by Capacity Constraints Bound the Scope of Intelligence

    Publicat: 03.05.2025
18. Toward Efficient Exploration by Large Language Model Agents

    Publicat: 03.05.2025
19. Getting More Juice Out of the SFT Data: Reward Learning from Human Demonstration Improves SFT

    Publicat: 02.05.2025
20. Self-Consuming Generative Models with Curated Data

    Publicat: 02.05.2025