Make synthetic trace block identities parent-sensitive

runs/frontier-code-trace-v0/experiment-card.md

@@ -228,6 +228,11 @@ mode-specific：
 - source 无 Qwen-aligned prompt/token IDs。raw canonical prompt 使用 GLM
   token contract，不能同时保持 Qwen token content 与 trace ISL；本 campaign
   采用 synthetic Qwen tokens 保持 length/hash/prefix shape，并降级内容 claim。
+- selected rho=0.0032 中有 1355 个可检查 parent links：tail rewrite
+  p50/p90/p95/p99/max=1/1/4/57/169 个 source blocks，说明 coder
+  `parent_chat_id` 不等价于 append-only prompt。source hash 序列作为 prefix
+  truth；synthetic content block 生成后再计算 parent-sensitive runtime
+  identities，避免“相同内容块出现在不同前缀后”造成 Frontier false hit。
 - 远端 Qwen3-30B `config.json` 的原生 position limit 是 40960
   (`rope_theta=1e6`,无 rope_scaling)。147456 profile smoke 在显式
   `VLLM_ALLOW_LONG_MAX_MODEL_LEN=1` 下成功；该 override 只支持

runs/frontier-s3-real-v0/remap_hash_blocks.py

@@ -257,17 +257,22 @@ def materialize(args: argparse.Namespace, *, tokenizer: Any | None = None) -> di
                     prompt_kind = "real_tokens"
                     real_prompt_requests += 1
                 else:
-                    block_ids = expand_hash_ids(
+                    content_block_ids = expand_hash_ids(
                         original_hashes,
                         isl,
                         source_block_size=source_block_size,
                     )
                     request_prompt = synthetic_tokens(
-                        block_ids,
+                        content_block_ids,
                         isl,
                         vocab_size=args.vocab_size,
                         token_offset=args.token_offset,
                     )
+                    records = token_block_identity_records(request_prompt)
+                    block_ids = [identity for identity, _ in records]
+                    for runtime_id, witness in records:
+                        previous = identity_to_witness.setdefault(runtime_id, witness)
+                        identity_collisions += int(previous != witness)
                     prompt_kind = "synthetic_missing_prompt_fallback"
                     synthetic_fallback_requests += 1
             parent = row.get("parent_chat_id")
@@ -417,8 +422,9 @@ def materialize(args: argparse.Namespace, *, tokenizer: Any | None = None) -> di
         "workload_mode": workload_mode,
         "mapping": (
             "real prompt tokens: BLAKE2b-128(parent runtime identity, exact 16-token block); "
-            "missing prompt fallback: expanded_id = zigzag(source_hash) * "
+            "missing prompt fallback: source hashes first define deterministic 16-token "
-            "subblocks_per_source + subblock_index"
+            "content blocks, then the same parent-sensitive BLAKE2b-128 runtime identity "
+            "contract is applied"
         ),
         "prompt_contract": {
             "real_prompt_requests": real_prompt_requests,
@@ -455,6 +461,7 @@ def materialize(args: argparse.Namespace, *, tokenizer: Any | None = None) -> di
             "encoding": "injective base-(vocab_size-token_offset), 16 little-endian digits",
             "same_16_block_hash_same_tokens": True,
             "different_16_block_hash_different_tokens": True,
+            "runtime_identity_is_parent_sensitive": True,
         },
         "parent_validation": {
             "enabled": args.validate_parents,

runs/frontier-s3-real-v0/test_remap_hash_blocks.py

@@ -59,6 +59,24 @@ class StrictRemapTest(unittest.TestCase):
         token_blocks_b = [tuple(token_block(value, vocab_size=151936)) for value in request_b]
         self.assertEqual(common_prefix_length(token_blocks_a, token_blocks_b), remapped_hits)
         self.assertEqual(len(set(token_blocks_a)), len(set(request_a)))
+        runtime_a = token_block_identities(
+            synthetic_tokens(request_a, 180, vocab_size=151936)
+        )
+        runtime_b = token_block_identities(
+            synthetic_tokens(request_b, 190, vocab_size=151936)
+        )
+        self.assertEqual(common_prefix_length(runtime_a, runtime_b), remapped_hits)
+
+    def test_synthetic_runtime_identity_depends_on_parent_prefix(self) -> None:
+        left_content = expand_hash_ids([11, 22], 128)
+        right_content = expand_hash_ids([33, 22], 128)
+        left = token_block_identities(
+            synthetic_tokens(left_content, 128, vocab_size=151936)
+        )
+        right = token_block_identities(
+            synthetic_tokens(right_content, 128, vocab_size=151936)
+        )
+        self.assertNotEqual(left[4], right[4])
 
     def test_mapping_rejects_non_strict_source_block_count(self) -> None:
         with self.assertRaisesRegex(ValueError, "strict 64-block contract"):