Best AI papers explained

550 Afleveringen

1. Front-Loading Reasoning: The Synergy between Pretraining and Post-Training Data

   Gepubliceerd: 18-10-2025
2. Representation-Based Exploration for Language Models: From Test-Time to Post-Training

   Gepubliceerd: 18-10-2025
3. The attacker moves second: stronger adaptive attacks bypass defenses against LLM jail- Breaks and prompt injections

   Gepubliceerd: 18-10-2025
4. When can in-context learning generalize out of task distribution?

   Gepubliceerd: 16-10-2025
5. The Art of Scaling Reinforcement Learning Compute for LLMs

   Gepubliceerd: 16-10-2025
6. A small number of samples can poison LLMs of any size

   Gepubliceerd: 16-10-2025
7. Dual Goal Representations

   Gepubliceerd: 14-10-2025
8. Welcome to the Era of Experience

   Gepubliceerd: 14-10-2025
9. Value Flows: Flow-Based Distributional Reinforcement Learning

   Gepubliceerd: 14-10-2025
10. Self-Adapting Language Models

    Gepubliceerd: 12-10-2025
11. The Markovian Thinker

    Gepubliceerd: 12-10-2025
12. Moloch’s Bargain: emergent misalignment when LLMs compete for audiences

    Gepubliceerd: 12-10-2025
13. Transformer Predictor Dynamics and Task Diversity

    Gepubliceerd: 11-10-2025
14. Base models know how to reason, thinking models learn when

    Gepubliceerd: 11-10-2025
15. Spectrum tuning: Post-training for distributional coverage and in-context steerability

    Gepubliceerd: 11-10-2025
16. Understanding Prompt Tuning and In-Context Learning via Meta-Learning

    Gepubliceerd: 11-10-2025
17. MLPs Learn In-Context on Regression and Classification tasks

    Gepubliceerd: 11-10-2025
18. Is Pre-Training Truly Better than Meta-Learning?

    Gepubliceerd: 11-10-2025
19. Agentic Context Engineering: Evolving Contexts for Self-Improving Language Models

    Gepubliceerd: 11-10-2025
20. Do LLMs Recognize Your Preferences? Evaluating Personalized Preference Following in LLMs

    Gepubliceerd: 9-10-2025