File size: 5,686 Bytes
0a5b3cd
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
{
  "schema": "mt11-high-heavy-l45-primary-selection/1",
  "status": "FINAL_FROZEN",
  "date": "2026-08-16",
  "selection": {
    "primary": "HIGH_HEAVY_L45",
    "downgraded": "MEDCONF_MED34_HIGH30_CENTERED_S2",
    "downgrade_evidence": {
      "fresh_correct": 29,
      "fresh_n": 240,
      "vs_fresh_base_delta_correct": 9,
      "question_cluster_bootstrap_95ci_delta_correct": [
        -3,
        22
      ],
      "significantly_above_fresh_base": false
    },
    "owner_override": {
      "applied": true,
      "prior_report_verdict": "DO_NOT_STOP_TRAINING_MIXED_OR_UNSTABLE",
      "reason": "The prior report verdict combined HIGH and MEDCONF. The owner had predeclared that fresh-seed reproduction of the 34-point result ends training; HIGH exactly reproduced 34/240 and remained significantly above Fresh Base."
    }
  },
  "high_complete_evidence": {
    "no_cherry_pick": true,
    "all_seeds_included": [
      42,
      43,
      44,
      45,
      46,
      47,
      48,
      49
    ],
    "pairing_note": "Different eval seeds are not sample-paired; comparisons use question aggregation. Combined health metrics are exhaustive sample totals.",
    "original_block": {
      "seeds": [
        42,
        43,
        44,
        45
      ],
      "n": 240,
      "correct": 34,
      "correct_by_seed": {
        "42": 8,
        "43": 9,
        "44": 6,
        "45": 11
      },
      "cap": {
        "n": 91,
        "pct": 37.917
      },
      "decoded_complete_box": {
        "n": 150,
        "pct": 62.5
      },
      "natural_stop": {
        "n": 149,
        "pct": 62.083
      }
    },
    "fresh_block": {
      "seeds": [
        46,
        47,
        48,
        49
      ],
      "n": 240,
      "correct": 34,
      "correct_by_seed": {
        "46": 8,
        "47": 9,
        "48": 9,
        "49": 8
      },
      "cap": {
        "n": 84,
        "pct": 35.0
      },
      "decoded_complete_box": {
        "n": 151,
        "pct": 62.917
      },
      "natural_stop": {
        "n": 156,
        "pct": 65.0
      },
      "vs_fresh_base": {
        "fresh_base_correct": 20,
        "fresh_base_n": 240,
        "delta_correct": 14,
        "question_cluster_bootstrap_95ci_delta_correct": [
          2,
          27
        ],
        "significantly_above": true
      }
    },
    "combined_block": {
      "seeds": [
        42,
        43,
        44,
        45,
        46,
        47,
        48,
        49
      ],
      "n": 480,
      "correct": {
        "n": 68,
        "pct": 14.167
      },
      "cap": {
        "n": 175,
        "pct": 36.458
      },
      "decoded_complete_box": {
        "n": 301,
        "pct": 62.708
      },
      "natural_stop": {
        "n": 305,
        "pct": 63.542
      }
    }
  },
  "frozen_identity": {
    "adapter_path": "/workspace/DDC-v3.1/runs/mt11_r2_384k96_execution_20260814/arms_90r6pp/HIGH_HEAVY_L45/train/checkpoint-update-00000024-tokens-000000409131/adapter",
    "adapter_sha256": "628a5b4c40584c524479b80773166f49b5c49d7223b34d4cf254064d6b65bd29",
    "base_path": "/workspace/MATH-NOTHINK-AIME-P1-v1.0/model/mn8_eos_r1_2/Qwen3-4B-Base-906bfd4",
    "tokenizer_path": "/workspace/MATH-NOTHINK-AIME-P1-v1.0/model/mn8_eos_r1_2/Qwen3-4B-Base-906bfd4",
    "dynamic_lora": true,
    "base_adapter_separate": true,
    "merge": false,
    "bake": false
  },
  "standard_think_evalscope_contract": {
    "mode": "think",
    "enable_thinking": true,
    "framework": "EvalScope",
    "score_authority": "EvalScope reviews",
    "suites": [
      "AIME24",
      "AIME25"
    ],
    "independent_questions": 60,
    "samples_per_four_seed_block": 240,
    "sampling": {
      "temperature": 0.6,
      "top_p": 0.95,
      "top_k": 20
    },
    "max_model_len": 32768,
    "reserve_tokens": 64,
    "max_new_tokens_rule": "32768 - exact_upstream_tokenize_prompt_count - 64",
    "statistics": "question-clustered; do not treat 240 samples as 240 independent questions",
    "serving": "vLLM dynamic LoRA mount; no merge or bake"
  },
  "route_freeze": {
    "stop_all_new_mt11_training": true,
    "short_cot": "FAIL",
    "med34_score_uplift_r1_family_gate": "FAIL",
    "automatic_continuation_forbidden": true,
    "automatic_arm_expansion_forbidden": true,
    "automatic_warm_start_forbidden": true,
    "automatic_lora_combination_forbidden": true,
    "historical_result_cards_modified": false,
    "training_started_by_this_action": false,
    "evaluation_started_by_this_action": false
  },
  "mn9_crossprotocol_eval": {
    "approved_eval_only": true,
    "must_not_stop_or_interfere": true,
    "may_not_trigger_training_automatically": true,
    "may_not_trigger_lora_combination_automatically": true,
    "separate_approval_required_for_any_followup": true
  },
  "supervisor_read_only_check": {
    "mt11_training_running": false,
    "action_taken": "NONE",
    "running_approved_eval_only_services": [
      "ddc-mt11-mn9-crossprotocol-controller",
      "ddc-mt11-mn9-crossprotocol-eval-disk-guard",
      "ddc-mt11-mn9-crossprotocol-eval-gpu0",
      "ddc-mt11-mn9-crossprotocol-eval-gpu1",
      "ddc-mt11-mn9-crossprotocol-eval-gpu2",
      "ddc-mt11-mn9-crossprotocol-eval-gpu3"
    ]
  },
  "authoritative_sources": [
    "/workspace/DDC-v3.1/runs/mt11_two34_freshseed_reval_20260815/MT11_TWO34_FRESHSEED_REVAL_FINAL.md",
    "/workspace/DDC-v3.1/runs/mt11_two34_freshseed_reval_20260815/MT11_TWO34_FRESHSEED_REVAL_FINAL.json",
    "/workspace/DDC-v3.1/runs/mt11_r2_384k96_execution_20260814/RESULT_CARD_HIGH_HEAVY_L45.json",
    "/workspace/DDC-v3.1/runs/MED34_SCORE_UPLIFT_CANARY_V1R1_SHARED_PREFIX/MED34_SCORE_UPLIFT_REPORT.md"
  ]
}