Leon Guertler

SPIRAL: Self-Play on Zero-Sum Games Incentivizes Reasoning via Multi-Agent Multi-Turn Reinforcement Learning

TeLLMe what you see: Using LLMs to Explain Neurons in Vision Models

Adding 32 Parameters to a LLM can improve fine-tuned classification performance by up to 1.5-6 percentage points

NoiseOut: Learning to Gate Improves Robustness in Deep Neural Networks