RiskLab
Toolkit · ACL 2026 Demo
A controlled multi-agent framework for probing emergent social risks in LLM agent collectives, specified as topology–environment–protocol–agent–task quintuples.
Toolkits & Datasets
Research Artifacts
Open-source toolkits, benchmarks, and datasets from our work on trustworthy and aligned foundation models.
ProbeLLM
Toolkit · ICML 2026
An automated probing framework that discovers structured LLM failure modes via hierarchical Monte Carlo Tree Search, tool-augmented test generation, and failure-aware clustering.
IntraAI
Software
Built around three goals: accessible AI for all, adaptive learning and growth, and trust with collective understanding.
Software — release soon
EmoNest
Software · NeurIPS 2025 Creative AI
A generative AI framework for creating interactive, emotionally adaptive storytelling games.
Demo — will release in Dec. at conference
ValueLence
Dashboard
The first unified platform for dynamic, fine-grained value probing of LLMs — value curation, probe generation, response collection, and multi-dimensional evaluation.
SDE-Harness
Toolkit
Scientific Discovery Evaluation — an extensible framework designed to accelerate AI-powered scientific discovery.
ChemOrch
Toolkit · NeurIPS 2025
Intelligent task orchestration for chemical research — turns a chemical task into high-quality instruction–response pairs.
TrustEval
Toolkit · NAACL 2025 Demo
A modular, extensible toolkit for trust evaluation of generative foundation models across safety, fairness, robustness, privacy, and more.
DataGen
Toolkit · ICLR 2025
An LLM-powered framework for generating diverse, accurate, and highly controllable text datasets.
TrustLLM
Toolkit · ICML 2024
A Python package that helps you assess the trustworthiness of your LLM more quickly.
MetaTool
Dataset · ICLR 2024
A benchmark for evaluating whether LLMs have tool-usage awareness and can correctly choose the right tool.