Commit History

SpreadsheetBench: SkillOpt mode=multi — 30-turn code repair with execution feedback, 600 s task budget, training-only gold cell feedback
f24ba1b
verified

emrekuruu commited on

SpreadsheetBench: actor prompts aligned verbatim with the SkillOpt codegen harness
a15f83f
verified

emrekuruu commited on

Argument descriptions carry format only; environment description is the single home of world rules
adaffb2
verified

emrekuruu commited on

fix: prevent stalled SpreadsheetBench sessions
81a00cc
verified

emrekuruu commited on

feat: expose native typed action tools
2869fbc
verified

emrekuruu commited on

Preserve official SpreadsheetBench grading positions
9426326
verified

emrekuruu commited on

Stage isolated SpreadsheetBench concurrency candidate
3f44ff2
verified

emrekuruu commited on

Isolate runner identities and cap BLAS threads
1fa02d1
verified

emrekuruu commited on

Freeze public SpreadsheetBench task identities
ec813f6
verified

emrekuruu commited on

Remove generated bytecode from Space
e250eb8
verified

emrekuruu commited on

Deploy standalone SpreadsheetBench OpenEnv
b4a5305
verified

emrekuruu commited on