violetxi/tb2-eval-qwen3-8b-base
tb2-eval-qwen3-8b-base: Terminal-Bench 2.0 eval results - Qwen3-8B base: Harbor dataset on Hugging Face. Compact Harbor eval results for Qwen/Qwen3-8B on Terminal-Bench 2.0 using the terminus-2 agent harness. Each dataset split is one checkpoint step. Rows are Harbor trial directories and include…