Best AI papers explained

178 Jaksot

1. Asymptotic Safety Guarantees Based On Scalable Oversight

   Julkaistiin: 6.5.2025
2. What Makes a Reward Model a Good Teacher? An Optimization Perspective

   Julkaistiin: 6.5.2025
3. Towards Guaranteed Safe AI: A Framework for Ensuring Robust and Reliable AI Systems

   Julkaistiin: 6.5.2025
4. Identifiable Steering via Sparse Autoencoding of Multi-Concept Shifts

   Julkaistiin: 6.5.2025
5. You Are What You Eat - AI Alignment Requires Understanding How Data Shapes Structure and Generalisation

   Julkaistiin: 6.5.2025
6. Interplay of LLMs in Information Retrieval Evaluation

   Julkaistiin: 3.5.2025
7. Trade-Offs Between Tasks Induced by Capacity Constraints Bound the Scope of Intelligence

   Julkaistiin: 3.5.2025
8. Toward Efficient Exploration by Large Language Model Agents

   Julkaistiin: 3.5.2025
9. Getting More Juice Out of the SFT Data: Reward Learning from Human Demonstration Improves SFT

   Julkaistiin: 2.5.2025
10. Self-Consuming Generative Models with Curated Data

    Julkaistiin: 2.5.2025
11. Bootstrapping Language Models with DPO Implicit Rewards

    Julkaistiin: 2.5.2025
12. DeepSeek-Prover-V2: Advancing Formal Reasoning

    Julkaistiin: 1.5.2025
13. THINKPRM: Data-Efficient Process Reward Models

    Julkaistiin: 1.5.2025
14. Societal Frameworks and LLM Alignment

    Julkaistiin: 29.4.2025
15. Risks from Multi-Agent Advanced AI

    Julkaistiin: 29.4.2025
16. Causality-Aware Alignment for Large Language Model Debiasing

    Julkaistiin: 29.4.2025
17. Reward Models Evaluate Consistency, Not Causality

    Julkaistiin: 28.4.2025
18. Causal Rewards for Large Language Model Alignment

    Julkaistiin: 28.4.2025
19. Sycophancy to subterfuge: Investigating reward-tampering in large language models

    Julkaistiin: 28.4.2025
20. Bidirectional AI Alignment

    Julkaistiin: 28.4.2025