Best AI papers explained

427 Episodes

1. Reward Shaping from Confounded Offline Data

   Published: 5/25/2025
2. Trajectory Bellman Residual Minimization: A Simple Value-Based Method for LLM Reasoning

   Published: 5/25/2025
3. Understanding Best-of-N Language Model Alignment

   Published: 5/25/2025
4. Maximizing Acquisition Functions for Bayesian Optimization - and its relation to Gradient Descent

   Published: 5/24/2025
5. Bayesian Prompt Ensembles: Model Uncertainty Estimation for Black-Box Large Language Models

   Published: 5/24/2025
6. Prompting Strategies for Enabling Large Language Models to Infer Causation from Correlation

   Published: 5/24/2025
7. The Parallel Knowledge Gradient Method for Batch Bayesian Optimization

   Published: 5/24/2025
8. FunBO: Discovering Acquisition Functions for Bayesian Optimization with FunSearch

   Published: 5/24/2025
9. Automated Social Science: A Structural Causal Model-Based Approach

   Published: 5/24/2025
10. Causal Interpretation of Transformer Self-Attention

    Published: 5/24/2025
11. A Causal World Model Underlying Next Token Prediction: Exploring GPT in a Controlled Environment

    Published: 5/24/2025
12. Trace is the Next AutoDiff: Generative Optimization with Rich Feedback, Execution Traces, and LLMs

    Published: 5/24/2025
13. Adaptive Inference-Time Compute: LLMs Can Predict if They Can Do Better, Even Mid-Generation

    Published: 5/24/2025
14. Prompts from Reinforcement Learning (PRL)

    Published: 5/24/2025
15. Logits are All We Need to Adapt Closed Models

    Published: 5/24/2025
16. Large Language Models Are (Bayesian) Latent Variable Models: Explaining and Finding Good Demonstrations for In-Context Learning

    Published: 5/23/2025
17. Inference-Time Intervention: Eliciting Truthful Answers from a Language Model

    Published: 5/23/2025
18. From Decoding to Meta-Generation: Inference-time Algorithms for Large Language Models

    Published: 5/23/2025
19. LLM In-Context Learning as Kernel Regression

    Published: 5/23/2025
20. Personalizing LLMs via Decode-Time Human Preference Optimization

    Published: 5/23/2025