T-Eval

What is T-Eval?

T-Eval evaluates the tool utilization capabilities of LLMs and decomposing them into instruction following, planning, reasoning, retrieval, understanding, and…

Released

2024-01-15

Evaluates

智能体, agent, 任务执行, Task Execution, 官方自建, Official, 不支持, Unsupported

Openness

open

Importer

OpenCompass Hub

Reported scores

0

Source provenance

• Original evidence https://hub.opencompass.org.cn/dataset-detail/T-Eval

T-Eval paper, code and dataset

• Paper https://arxiv.org/abs/2312.14033
• Code https://github.com/open-compass/T-Eval
• Dataset https://huggingface.co/datasets/lovesnowbest/T-Eval

Which sources cite T-Eval?

• T-Eval

Related 智能体, agent, 任务执行, Task Execution, 官方自建, Official, 不支持, Unsupported benchmarks

• A-Bench
• A-OKVQA
• AbsPyramid
• ActionAtlas
• Ada-LEval
• AECBench
• AFQMC
• AgentBoard