Qwen3-4B checkpoints whose chat template always preserves <think> reasoning across multi-turn history, with no nonthinking mode.
Ethan Ewer
eewer
AI & ML interests
None yet
Recent Activity
updated a model 3 days ago
eewer/qwen3.5-35b-a3b-projectbro-rl-65k-iter99 published a model 3 days ago
eewer/qwen3.5-35b-a3b-projectbro-rl-65k-iter99 updated a model 4 days ago
eewer/qwen3.5-35b-a3b-glm52-multi-harness-65k-lr5e6-step808Organizations
None yet
models 71
eewer/qwen3.5-35b-a3b-projectbro-rl-65k-iter99
Text Generation • 35B • Updated • 20
eewer/qwen3.5-35b-a3b-glm52-multi-harness-65k-lr5e6-step808
Text Generation • 35B • Updated • 15 • 1
eewer/qwen3.5-35b-a3b-glm52-multi-harness-65k-lr1e5-step808
Text Generation • 35B • Updated • 19
eewer/qwen3.5-35b-a3b-base-terminal_swe_general-agent-ablation-step7176
Text Generation • 35B • Updated • 24 • 1
eewer/qwen3.5-35b-a3b-base-terminal_swe-agent-ablation-step7073
Text Generation • 35B • Updated • 22
eewer/qwen3.5-35b-a3b-base-terminal_swe_general-agent-step7176
Text Generation • 35B • Updated • 13
eewer/qwen3.5-35b-a3b-base-terminal_swe-agent-step7073
Text Generation • 35B • Updated • 22
eewer/qwen3.5-35b-a3b-base-terminal-agent-step2961
Text Generation • 35B • Updated • 17
eewer/qwen3.5-35b-a3b-base-terminal-agent-step5271
Text Generation • 35B • Updated • 20
eewer/qwen3.5-35b-a3b-base-terminal_swe_general-agent-step2500
Text Generation • 35B • Updated • 63
datasets 7
eewer/mixed-sft-openai-tools-qwen3-areal-diverse
Updated • 58
eewer/skill-based-medium-terminus2-sft
Updated • 66
eewer/qwen3-4b-thinking-sft-v54-raw2030-strictpassed-processed
Preview • Updated • 76
eewer/swerebench-traces-raw-source-verification-enhanced-20260617
Viewer • Updated • 1 • 24
eewer/agent-dataset
Viewer • Updated • 869 • 10
eewer/MMLU-Pro-engineering-outputs
Viewer • Updated • 25.6k • 26
eewer/deepseek-MMLU-Pro-engineering-outputs
Viewer • Updated • 8.69k • 44