{"data":{"kind":"file","path":"README.md","version_id":"jhnc2g0d0pqctb82aaaavq5h","entry":{"name":"README.md","path":"README.md","is_directory":false,"size":791,"modified_at":"2026-04-29T10:50:44.655000","content_hash":"16528af892e2700585842a3b956eb2cc2721cb10bde8c77cd9d8eaed4181fbd8"},"entries":[],"content":"# phase-retrieval-multiturn\n\n3-turn phase retrieval with magnitude-residual feedback between turns.\n\nVerifiable Labs Scientific-RL environment. Published as a thin wrapper around the monorepo at https://github.com/stelioszach03/verifiable-labs-envs — the wrapper pulls the monorepo as a Git dependency so the full source of truth (rewards, forward operators, LLM adapter, conformal calibration) stays in one place.\n\n## Install\n\n```bash\nprime env install verifiable-labs/phase-retrieval-multiturn\n```\n\n## Use\n\n```python\nfrom verifiers import load_environment    # or Prime SDK equivalent\nenv = load_environment(\"phase-retrieval-multiturn\")\nout = env.run_baseline(seed=0)\nprint(out[\"reward\"])\n```\n\nSee the monorepo README + docs for the reward spec, contamination story, and benchmark data.\n","encoding":"utf-8","truncated":false,"total_bytes":791},"status":null}