Best AI papers explained

527 Episodes

1. Bayesian Scaling Laws for In-Context Learning

   Published: 15/05/2025
2. Posterior Mean Matching Generative Modeling

   Published: 15/05/2025
3. Can Generative AI Solve Your In-Context Learning Problem? A Martingale Perspective

   Published: 15/05/2025
4. Dynamic Search for Inference-Time Alignment in Diffusion Models

   Published: 15/05/2025
5. Is In-Context Learning in Large Language Models Bayesian? A Martingale Perspective

   Published: 12/05/2025
6. Leaked Claude Sonnet 3.7 System Instruction tuning

   Published: 12/05/2025
7. Converging Predictions with Shared Information

   Published: 11/05/2025
8. Test-Time Alignment Via Hypothesis Reweighting

   Published: 11/05/2025
9. Rethinking Diverse Human Preference Learning through Principal Component Analysis

   Published: 11/05/2025
10. Active Statistical Inference

    Published: 10/05/2025
11. Data Mixture Optimization: A Multi-fidelity Multi-scale Bayesian Framework

    Published: 10/05/2025
12. AI-Powered Bayesian Inference

    Published: 10/05/2025
13. Can Unconfident LLM Annotations Be Used for Confident Conclusions?

    Published: 9/05/2025
14. Predictions as Surrogates: Revisiting Surrogate Outcomes in the Age of AI

    Published: 9/05/2025
15. Learn then Test: Calibrating Predictive Algorithms to Achieve Risk Control

    Published: 9/05/2025
16. How to Evaluate Reward Models for RLHF

    Published: 9/05/2025
17. LLMs as Judges: Survey of Evaluation Methods

    Published: 9/05/2025
18. The Alternative Annotator Test for LLM-as-a-Judge: How to Statistically Justify Replacing Human Annotators with LLMs

    Published: 9/05/2025
19. Limits to scalable evaluation at the frontier: LLM as Judge won’t beat twice the data

    Published: 9/05/2025
20. Stratified Prediction-Powered Inference for Hybrid Language Model Evaluation

    Published: 9/05/2025