The Arbiter Agent: Continually Monitoring Multi-Agent Conversations to Detect Emergent Misalignment
Paper • 2606.10747 • Published • 13
Lukas Galke Poech
lgalke
AI & ML interests
LLM interpretability, agentic/multi-agent safety
Recent Activity
authored a paper 8 days ago
The Arbiter Agent: Continually Monitoring Multi-Agent Conversations to Detect Emergent Misalignment authored a paper 14 days ago
BrainSurgery: Reproducible and Reliable Declarative Weight Manipulations for Model Editing and UpcyclingOrganizations
authored a
paper 8 days ago
upvoted a paper 9 days ago
authored a
paper 14 days ago
upvoted a paper 14 days ago
authored 10
papers 17 days ago
DeToNATION: Decoupled Torch Network-Aware Training on Interlinked Online Nodes
Paper • 2502.06728 • Published
Are We Really Making Much Progress in Text Classification? A Comparative Review
Paper • 2204.03954 • Published
Efficient Continual Learning for Small Language Models with a Discrete Key-Value Bottleneck
Paper • 2412.08528 • Published
FlexMoRE: A Flexible Mixture of Rank-heterogeneous Experts for Efficient Federatedly-trained Large Language Models
Paper • 2602.08818 • Published • 2
DaLA: Danish Linguistic Acceptability Evaluation Guided by Real World Errors
Paper • 2512.04799 • Published
SommBench: Assessing Sommelier Expertise of Language Models
Paper • 2603.12117 • Published • 1
Training Language Models to Use Prolog as a Tool
Paper • 2512.07407 • Published
The Moltbook Files: A Harmless Slopocalypse or Humanity's Last Experiment
Paper • 2605.07462 • Published • 3
Emergent Languages in Populations of Language Model Agents: From Token Efficiency to Oversight Evasion
Paper • 2605.31170 • Published • 12
LLMs Can Leak Training Data But Do They Want To? A Propensity-Aware Evaluation of Memorization in LLMs
Paper • 2606.06286 • Published • 8
upvoted a paper 17 days ago
upvoted a paper 20 days ago
liked a
model 27 days ago
authored a
paper 27 days ago
upvoted a paper 29 days ago
updated a
collection about 2 months ago