Renumber the pipeline substeps to match the paper (5-1 becomes 5a)

prompts/ and runs/ move by git mv; zero padding dropped; the arbitration and LLM-merge rows in the cost ledger carry no new name -- their step column reads 已廢棄 with the original name kept in a new last column. Archived meta.json files and past decision-log entries keep the names they were written with. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-18 23:37:32 +08:00
parent e3cca7e538
commit 85d776da8b
83 changed files with 144 additions and 114 deletions
@@ -0,0 +1,25 @@
+{
+  "script_version": "run_llm.py 3.1.0",
+  "model": "claude-fable-5",
+  "temperature": null,
+  "thinking": null,
+  "max_tokens": 16000,
+  "prompt_path": "../prompts/05-04-annotate.md",
+  "prompt_sha256": "c2312ba2a612e026dd9a616b84121aba8a3ab8e42abd49a1378857ad4b2630a1",
+  "input_path": "instance/llm-input-annotate-rescue.jsonl",
+  "input_sha256": "ebb94073524913bd519ddbf92b6e920f538bd31753600c94f4eb8ff3a1ee789d",
+  "item_count": 2,
+  "dry_run": false,
+  "started_at": "2026-08-15T19:09:40+08:00",
+  "batch": {
+    "batch_id": "msgbatch_01YEdroZ51b8btPfG1aqPFCD",
+    "submitted_at": "2026-08-15T19:09:41+08:00",
+    "ended_at": "2026-08-15T11:12:27.389327+00:00"
+  },
+  "usage": {
+    "cache_creation_input_tokens": 0,
+    "cache_read_input_tokens": 0,
+    "input_tokens": 17802,
+    "output_tokens": 11601
+  }
+}
@@ -0,0 +1,2 @@
+{"id": "song-177", "text": "[\"M1\", \"X2\", \"X4\", \"X5\", \"X8\", \"F14\", \"X16\"]", "stop_reason": "end_turn", "usage": {"cache_creation": {"ephemeral_1h_input_tokens": 0, "ephemeral_5m_input_tokens": 0}, "cache_creation_input_tokens": 0, "cache_read_input_tokens": 0, "inference_geo": "global", "input_tokens": 8900, "output_tokens": 3525, "output_tokens_details": {"thinking_tokens": 3494}, "service_tier": "batch"}}
+{"id": "song-199", "text": "[\"X5\", \"X8\", \"X9\", \"F3\", \"X12\", \"X14\", \"X15\", \"X16\"]", "stop_reason": "end_turn", "usage": {"cache_creation": {"ephemeral_1h_input_tokens": 0, "ephemeral_5m_input_tokens": 0}, "cache_creation_input_tokens": 0, "cache_read_input_tokens": 0, "inference_geo": "global", "input_tokens": 8902, "output_tokens": 8076, "output_tokens_details": {"thinking_tokens": 8041}, "service_tier": "batch"}}
@@ -0,0 +1,36 @@
+You are judging which patterns of feminist problems operate
+in one song, from the song's consolidated reading report.
+
+Input: a JSON object with the report and the patterns — the
+complete set to judge; use these and no others:
+
+{
+  "report": "the consolidated reading report of the song",
+  "patterns": [
+    {"id": "the-pattern-id",
+     "name": "the pattern name",
+     "description": "how the pattern operates"}
+  ]
+}
+
+Task: list every given pattern whose described mechanism
+operates in the problems the report states.
+
+Rules:
+
+- Judge only from the given report; do not use any knowledge
+  of the song beyond it.
+- A song may match any number of patterns, including none.
+- A pattern matches only when the report states a problem
+  working in the form the pattern describes; a shared topic
+  or vocabulary alone is not a match.
+- Use only the given pattern ids, spelled exactly as given.
+
+Do not wrap the output in a Markdown code fence.
+The output must be strictly valid JSON.
+
+Output a single JSON array of the ids of the matching
+patterns (an empty array when none matches), and nothing
+else:
+
+["the-pattern-id", "another-pattern-id"]