{"data":{"kind":"file","path":"README.md","version_id":"u9lun1egls0xp5oydmu8jaeu","entry":{"name":"README.md","path":"README.md","is_directory":false,"size":1772,"modified_at":"2025-10-31T22:54:26.072000","content_hash":"98f463a46051fa098c1fde47e25fbd424f67ae1f75a6aefca5424fef61ececf4"},"entries":[],"content":"# gsm8k\n\n### Overview\n- **Environment ID**: `gsm8k`\n- **Short description**: Single-turn GSM8K math word problems with boxed numeric answers and CoT.\n- **Tags**: math, gsm8k, single-turn, think, boxed-answer\n\n### Datasets\n- **Primary dataset(s)**: `gsm8k` train (train) and test (eval) via `load_example_dataset`\n- **Source links**: Uses the example loader in `verifiers.utils.data_utils`\n- **Split sizes**: Configurable via args; defaults to full train/test\n\n### Task\n- **Type**: single-turn\n- **Parser**: `ThinkParser` with boxed answer extraction\n- **Rubric overview**: Exact match on parsed boxed answer; optional format check\n\n### Quickstart\nRun an evaluation with default settings:\n\n```bash\nuv run vf-eval gsm8k\n```\n\nConfigure model and sampling:\n\n```bash\nuv run vf-eval gsm8k \\\n  -m gpt-4.1-mini \\\n  -n 20 -r 3 -t 1024 -T 0.7 \\\n  -a '{\"num_train_examples\": -1, \"num_eval_examples\": -1}'\n```\n\nNotes:\n- Use `-a` / `--env-args` to pass environment-specific configuration as a JSON object.\n- Reports are written under `./environments/gsm8k/reports/` and auto-embedded below.\n\n### Environment Arguments\n| Arg | Type | Default | Description |\n| --- | ---- | ------- | ----------- |\n| `num_train_examples` | int | `-1` | Limit training set size (`-1` for all) |\n| `num_eval_examples` | int | `-1` | Limit eval set size (`-1` for all) |\n\n### Metrics\n| Metric | Meaning |\n| ------ | ------- |\n| `reward` | 1.0 if parsed boxed answer equals target, else 0.0 |\n| `format_reward` | Adherence to `<think>` + boxed `\\boxed{...}` format |\n\n## Evaluation Reports\n\n<!-- Do not edit below this line. Content is auto-generated. -->\n<!-- vf:begin:reports -->\n<p>No reports found. Run <code>uv run vf-eval gsm8k -a '{\"key\": \"value\"}'</code> to generate one.</p>\n<!-- vf:end:reports -->\n","encoding":"utf-8","truncated":false,"total_bytes":1772},"status":null}