rolemrc:
  resources_servers:
    rolemrc:
      entrypoint: app.py
      domain: instruction_following
      verified: false
      description: RoleMRC role-play MRC scored with reference metrics (ROUGE/BLEU/METEOR/BERTScore); ROUGE-L is the reward.
      value: Improve role-play instruction following and knowledge-grounded responses.
      mode: reference
      # BERTScore is on by default for parity with the upstream benchmark; it
      # downloads a roberta-large checkpoint on first use. Set false (and drop
      # bert-score from requirements.txt) for a lightweight reward signal.
      include_bertscore: true

rolemrc_simple_agent:
  responses_api_agents:
    simple_agent:
      entrypoint: app.py
      resources_server:
        type: resources_servers
        name: rolemrc
      model_server:
        type: responses_api_models
        name: policy_model
      datasets:
      - name: example
        type: example
        jsonl_fpath: resources_servers/rolemrc/data/example.jsonl
        license: Creative Commons Attribution 4.0 International
      # Build with `python resources_servers/rolemrc/prepare_rolemrc.py`.
      - name: test
        type: validation
        jsonl_fpath: resources_servers/rolemrc/data/test.jsonl
        huggingface_identifier:
          repo_id: Junrulu/RoleMRC
          artifact_fpath: roleMRC_test.jsonl
        license: Creative Commons Attribution 4.0 International