llm-as-a-verifier 3

LLM-as-a-Verifier for dsh with the Best-of-N conversation mode built in: rank N candidates with a fine-grained verifier (expected grade over the logprob distribution), on demand via the verify tool or automatically on every turn of a Best-of-N session. In

多候选模式插件（test-time scaling）：🐋 悬浮球面板配置（启用/候选数/第二意见/验证标准，仅当前会话生效），模型自动对任务执行「并行候选→验证选优」流水线，无需任何特殊指令前缀。

Best-of-N sampling with an LLM-as-a-Verifier judge for DeepSeek Harness: sample N candidates per model call, score them with fine-grained logprob verification, and stream back the tournament winner.