healthbench_rubric: resources_servers: healthbench_rubric: entrypoint: app.py judge_model_server: type: responses_api_models name: judge_model judge_chat_completion_params: messages: [] temperature: 0.5 max_tokens: 2048 judge_endpoint_max_concurrency: 64 judge_max_attempts: 3 judge_timeout_seconds: 120 length_adjustment_center: null length_adjustment_penalty_per_500_chars: null reward_metric: overall_score domain: knowledge verified: false description: HealthBench-style clinical conversations scored with rubric LLM judge value: Evaluate medical response quality against private rubric criteria healthbench_rubric_simple_agent: responses_api_agents: simple_agent: entrypoint: app.py resources_server: type: resources_servers name: healthbench_rubric model_server: type: responses_api_models name: policy_model datasets: - name: example type: example jsonl_fpath: resources_servers/healthbench_rubric/data/example.jsonl healthbench_rubric_environment_server: environment_servers: single_agent_turn_legacy: entrypoint: app.py agent_server: type: responses_api_agents name: healthbench_rubric_simple_agent resources_server: type: resources_servers name: healthbench_rubric resources_tool_transports: - direct_http cleanup_timeout_seconds: 180