🌍 AppWorld: A Controllable World of Apps and People for Benchmarking Function Calling and Interactive Coding Agent, ACL'24 Best Resource Paper.

Separation of planning concerns in ReAct-style LLM agents. Planner fine-tuning on synthetic trajectories.

a suite of finetuned LLMs for atomically precise function calling 🧪

[ICML DMLR 2024] Repo that contains code for the paper titled: "Class-Level Code Generation from Natural Language Using Iterative, Tool-Enhanced Reasoning over Repository".

Benchmarking Large Language Models for Materials Science Tools

Code Base for Multi-Tool Usage for black box api access

🌍 Leaderboard Repository for "AppWorld: A Controllable World of Apps and People for Benchmarking Interactive Coding Agent", ACL2024

An autonomous research agent that synthesizes detailed reports by combining dynamic planning, unsupervised topic discovery (PCA/KMeans) HyDe (hypothetical doc embeddings), and a self-correcting reflexion loop.

An advanced conversational AI system with multi-model support, RAG capabilities, dynamic tool usage, and persistent memory.

A realistic OpenEnv environment for training AI agents to perform enterprise email triage across multi-email inbox workflows, with structured actions, tool usage, and reward shaping, built for the Scaler x Meta PyTorch Hackathon.

A self-improving coding agent that learns from its own mistakes.