Tau2
Active
Evaluating Conversational Agents in a Dual-Control Environment
- Publisher
- Sierra
- Domain
- Assistants
- License
- mit
- Published
- Dec 2025
- Notable for
- Benchmark for evaluating Assistants.
Cite
Notes
Only stored in your browser.
Related tools
4Implementations, trainers, datasets and scaffolds linked to this eval.
FAQ
- What is Tau2?
- Evaluating Conversational Agents in a Dual-Control Environment
- How can a model improve its Tau2 score?
- Tools linked to Tau2 on Sophon include Tau2 Bench (Jannik), Tau2 Synth (Prime), Tau2 Bench (Prime Intellect), Tau2 Bench (Will) - RL environments, datasets, and scaffolds that target this eval.
- What license is Tau2 under?
- Tau2 is available under mit.