Author

Dipika Khullar

2 papers indexed here

We haven’t gathered this author’s papers yet. Follow them and we’ll fetch their work.

Not the right person? Other researchers publish under this name.

#artificial intelligence Nov 2025

Co-Evolving Agents: Learning from Failures as Hard Negatives

A co-evolving framework that couples a target agent's self-improvement with a failure agent trained to generate hard negatives to demonstrate the value of learning to generate and use informative failures for agent self-improvement.

Yeonsung Jung, Trilok Padhi, S. Shaham et al. · arXiv.org · 8 citations

#artificial intelligence Jul 2026

Overthinking: Amplifying Reasoning Weights to Extract Learned Secrets

Overthinking is introduced, the process of using reasoning task vectors to amplify the propensity to think out loud of reasoning models, and reasoning amplification may surface secrets or unintended behaviors acquired during training up to $10\times$ more frequently than the original reasoning model.

Jack Hopkins, Dipika Khullar, Fabien Roger · arXiv.org · 1 citation