PPaperPicks

Christian Schröder de Witt

14 papers at tracked venues · 11 at CORE A* · active 20242025

Venues

Frequent coauthors

Papers

  1. Efficient Dictionary Learning with Switch Sparse Autoencoders
  2. Fundamental Limitations in Pointwise Defences of LLM Finetuning APIs
  3. Mixture of Experts Made Intrinsically Interpretable
  4. Multi-Agent Security Tax: Trading Off Security and Collaboration Capabilities in Multi-Agent Systems
  5. REAL: Benchmarking Autonomous Agents on Deterministic Simulations of Real Websites
  6. Bayesian Exploration Networks
  7. Computing Low-Entropy Couplings for Large-Support Distributions
  8. Illusory Attacks: Information-theoretic detectability matters in adversarial attacks
  9. JaxMARL: Multi-Agent RL Environments and Algorithms in JAX
  10. JaxMARL: Multi-Agent RL Environments and Algorithms in JAX
  11. Position: Near to Mid-term Risks and Opportunities of Open-Source Generative AI
  12. Rethinking Out-of-Distribution Detection for Reinforcement Learning: Advancing Methods for Evaluation and Detection
  13. Secret Collusion among AI Agents: Multi-Agent Deception via Steganography
  14. Unelicitable Backdoors via Cryptographic Transformer Circuits