Karthik Narasimhan

Introducing a framework for advanced text generation using language models.

Language agents merge AI and decision-making for real-world tasks.

This paper discusses how AI can learn from human language interactions.

SWE-agent improves LM agents' performance in software engineering tasks with a specialized interface.

A fresh approach to assess persona agents using language models.

New models improve content moderation for AI interactions, ensuring safety and reducing harm.

This framework improves lie detection in conversations during Diplomacy games.