J. Rosser

According to our database, J. Rosser authored at least 6 papers between 2025 and 2026.

Collaborative distances:

• Dijkstra number of five.
• Erdős number of four.

2026

Gradient Atoms: Unsupervised Discovery, Attribution and Steering of Model Behaviors via Sparse Decomposition of Training Gradients.

J. Rosser

CoRR, March, 2026

Infusion: Shaping Model Behavior by Editing Training Data via Influence Functions.

J. Rosser, Robert Kirk, Edward Grefenstette, Jakob N. Foerster, Laura Ruis

CoRR, February, 2026

2025

Mapping Faithful Reasoning in Language Models.

Jiazheng Li, Andreas Damianou, J. Rosser, José Luis Redondo García, Konstantina Palla

CoRR, October, 2025

Stream: Scaling up Mechanistic Interpretability to Long Context in LLMs via Sparse Attention.

J. Rosser, José Luis Redondo García, Gustavo Penha, Konstantina Palla, Hugues Bouchard

CoRR, October, 2025

AgentBreeder: Mitigating the AI Safety Impact of Multi-Agent Scaffolds.

J. Rosser, Jakob Nicolaus Foerster

CoRR, February, 2025

AgentBreeder: Mitigating the AI Safety Risks of Multi-Agent Scaffolds via Self-Improvement.

J. Rosser, Jakob N. Foerster

Proceedings of the Advances in Neural Information Processing Systems 38: Annual Conference on Neural Information Processing Systems 2025, 2025