diff --git a/evalharness/config/default.yaml b/evalharness/config/default.yaml index 47da330..58d8533 100644 --- a/evalharness/config/default.yaml +++ b/evalharness/config/default.yaml @@ -62,3 +62,7 @@ tau2_bench: max_tokens: 16384 env: tau2_official # agent 模式:官方引擎 max_turns: 200 # 一次完整电话流程(问候→验证→找单→政策→执行→确认) + +swe_bench_verified: + max_tokens: 32768 # patch 生成预算(diff 可能较长) + temperature: 0.0 # es 口径:确定性生成 patch