Genta Indra Winata commited on
Commit ·
b812ad6
1
Parent(s): c20499d
update README
Browse files
README.md
CHANGED
|
@@ -123,6 +123,8 @@ configs:
|
|
| 123 |
|
| 124 |
**T1-Bench** is a fully automated benchmark for evaluating the tool-calling capabilities of conversational AI agents across diverse service domains in task-oriented settings. The framework simulates end-to-end user–agent interactions without requiring human intervention at inference time, where a User Agent generates realistic customer utterances conditioned on predefined task goals and a tool-augmented Assistant Agent responds by invoking domain-specific tools/APIs and producing outputs grounded in tool results, prior conversational context, and domain-specific policies. Designed to capture the sequential and interactive nature of real-world service workflows across multi-domain scenarios, T1-Bench requires agents to maintain conversational state, reason over prior tool outputs, and execute multi-step operations such as search, filtering, booking, modification, and cancellation. Because all interactions are grounded in deterministic datasets and executable tools, the benchmark enables reproducible and fine-grained evaluation of agent behavior, tool-use decisions, and task completion performance.
|
| 125 |
|
|
|
|
|
|
|
| 126 |
## Results
|
| 127 |
|
| 128 |

|
|
@@ -145,5 +147,10 @@ The dataset is licensed under CC-BY-SA 4.0.
|
|
| 145 |
If you find this dataset useful, please cite the following work
|
| 146 |
|
| 147 |
```bibtex
|
| 148 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 149 |
```
|
|
|
|
| 123 |
|
| 124 |
**T1-Bench** is a fully automated benchmark for evaluating the tool-calling capabilities of conversational AI agents across diverse service domains in task-oriented settings. The framework simulates end-to-end user–agent interactions without requiring human intervention at inference time, where a User Agent generates realistic customer utterances conditioned on predefined task goals and a tool-augmented Assistant Agent responds by invoking domain-specific tools/APIs and producing outputs grounded in tool results, prior conversational context, and domain-specific policies. Designed to capture the sequential and interactive nature of real-world service workflows across multi-domain scenarios, T1-Bench requires agents to maintain conversational state, reason over prior tool outputs, and execute multi-step operations such as search, filtering, booking, modification, and cancellation. Because all interactions are grounded in deterministic datasets and executable tools, the benchmark enables reproducible and fine-grained evaluation of agent behavior, tool-use decisions, and task completion performance.
|
| 125 |
|
| 126 |
+
The paper is accepted at EMNLP Findings 2026. Check our .
|
| 127 |
+
|
| 128 |
## Results
|
| 129 |
|
| 130 |

|
|
|
|
| 147 |
If you find this dataset useful, please cite the following work
|
| 148 |
|
| 149 |
```bibtex
|
| 150 |
+
@article{winata2026t1,
|
| 151 |
+
title={T1-Bench: Benchmarking Multi-Scenario Agents in Real-World Domains},
|
| 152 |
+
author={Winata, Genta Indra and Chakraborty, Amartya and Lin, Yuzhen and Rao, Swasthi P and Siingh, Shikhhar and Lu, Houhan and Bathaee, Nadia and Hatwar, Sriharsha and Dashore, Paresh and Jain, Anmol and others},
|
| 153 |
+
journal={arXiv preprint arXiv:2606.11070},
|
| 154 |
+
year={2026}
|
| 155 |
+
}
|
| 156 |
```
|