Runner: G2 finds the FUNCTION statement after local classes; budget floor applied to all tasks

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_014aUaQeLnwbb1zTpN7kHeat
This commit is contained in:
Kral
2026-10-03 18:25:30 +02:00
parent 7351849a76
commit 92e30eb554
78 changed files with 1362 additions and 560 deletions

View File

@@ -1,16 +1,35 @@
{
"deepseek-v4.1-flash:cloud": {
"score": 30,
"parts": {
"total": 30,
"stop": "stopped",
"gap": {
"how": "empty report"
}
"_superseded": {
"deepseek-v4.1-flash:cloud@2026-10-03": {
"score": 30,
"parts": {
"total": 30,
"stop": "stopped",
"gap": {
"how": "empty report"
}
},
"hidden": "0/0",
"tool_calls": 0,
"seconds": 629.3,
"run_dir": "runs/emp/12019_G0169_llm_deepseek-v4.1-flash_cloud"
},
"hidden": "0/0",
"tool_calls": 0,
"seconds": 629.3,
"run_dir": "runs/emp/12019_G0169_llm_deepseek-v4.1-flash_cloud"
"deepseek-v4.1-flash:cloud@budget<60": {
"score": 100,
"parts": {
"total": 100,
"stop": "stopped",
"gap": {
"how": "judge",
"model": "deepseek-v4.1-flash:cloud",
"names_gap": true,
"reason": "The report explicitly identifies the missing precedence rule for members who are both students and under 25, matching the expected gap."
}
},
"hidden": "0/0",
"tool_calls": 0,
"seconds": 7.3,
"run_dir": "runs/emp/16009_G0169_llm_deepseek-v4.1-flash_cloud"
}
}
}

View File

@@ -27,8 +27,8 @@
]
},
"budget": {
"max_tool_calls": 40,
"max_activations": 10
"max_tool_calls": 60,
"max_activations": 15
},
"seed": [],
"contract": [