rl-rag/

hle-gpt-oss-120b-with-python-260222

Deep research agent evaluation on unknown.

	Results




	

Metric Value

pass@4 39.5%

avg@4 17.5%

Trajectory accuracy 17.4% (1860/10660)

Questions 1350

Trajectories 10660 (4 per question)

Avg tool calls 0.0

Full conversations ❌

	Model & Setup




	




	

Model unknown

Judge gpt-4o

Max tool calls 50

Temperature 0.7

Blocked domainsNone

	Tool Usage




	

Tool Calls %… See the full description on the dataset page: https://huggingface.co/datasets/rl-rag/hle-gpt-oss-120b-with-python-260222.