The </think> row edit that stops thinking-mode runaway. ThinkFix = abliterated + fix; Reasoning-Termination-Fix = stock weights + fix only.
Josh Bolding PRO
BoldingBuilds
P(doom) 1%
AI & ML interests
Independent builder working on local AI, reasoning models, and practical agents. I test models, experiment with small weight edits, and share results, GGUFs, and evaluation tools.
Recent Activity
repliedto their post about 17 hours ago
Abliterated thinking models have a habit: they decide what to say, then keep arguing with themselves until the token budget runs out. Qwen3.8-27B Abliterated ThinkFix is an uncensored Qwen3.8-27B with one output row edited so the model closes its thinking when it's ready to answer.
On 40 sensitive prompts the edit never saw, 8k output cap: 33 replies finished cleanly instead of 22, 5 hit the limit instead of 17, 61 minutes for the set instead of 75.
Agent use was checked four ways and the edit changes nothing there. MTP draft head kept. Q4_K_M to Q8_0, each tested after the edit. The card lists what the edit was fit on and what it does not do.
https://huggingface.co/BoldingBuilds/Qwen3.8-27B-Abliterated-ThinkFix-GGUF new activity 1 day ago
BoldingBuilds/Qwen3.8-27B-Abliterated-ThinkFix-GGUF:great release !!! updated a collection 1 day ago
MiMo V2.6 Distill Qwen 9B - AbliteratedOrganizations
Refusal-ablated GGUFs of XiaomiMiMo MiMo-V2.6-Distill-Qwen-9B.
Uncensored GGUF quants of GLM-5.3-Flash: our own expert-editing abliteration, and an imatrix requant of orcarouter's FP8 abliteration.
Abliterated prism-ml Ternary Bonsai 2 27B GGUFs. V2 PQ2_0+MTP is the current pick; V1 repos stay up. All need the prism llama.cpp fork.
-
BoldingBuilds/Ternary-Bonsai-2-27B-Abliterated-v2-PQ2_0-MTP-GGUF
Text Generation • 27B • Updated • 15.8k • 11 -
BoldingBuilds/Ternary-Bonsai-2-27B-Abliterated-PQ2_0-MTP-GGUF
Text Generation • 27B • Updated • 44.1k • 26 -
BoldingBuilds/Ternary-Bonsai-2-27B-Abliterated-PTQ1_0-GGUF
Text Generation • 27B • Updated • 75.5k • 71
ShimQuant GGUF builds of NVIDIA Nemotron 3.5 Lightning 30B-A3B, stock and abliterated.
The </think> row edit that stops thinking-mode runaway. ThinkFix = abliterated + fix; Reasoning-Termination-Fix = stock weights + fix only.
Abliterated prism-ml Ternary Bonsai 2 27B GGUFs. V2 PQ2_0+MTP is the current pick; V1 repos stay up. All need the prism llama.cpp fork.
-
BoldingBuilds/Ternary-Bonsai-2-27B-Abliterated-v2-PQ2_0-MTP-GGUF
Text Generation • 27B • Updated • 15.8k • 11 -
BoldingBuilds/Ternary-Bonsai-2-27B-Abliterated-PQ2_0-MTP-GGUF
Text Generation • 27B • Updated • 44.1k • 26 -
BoldingBuilds/Ternary-Bonsai-2-27B-Abliterated-PTQ1_0-GGUF
Text Generation • 27B • Updated • 75.5k • 71
Refusal-ablated GGUFs of XiaomiMiMo MiMo-V2.6-Distill-Qwen-9B.
ShimQuant GGUF builds of NVIDIA Nemotron 3.5 Lightning 30B-A3B, stock and abliterated.
Uncensored GGUF quants of GLM-5.3-Flash: our own expert-editing abliteration, and an imatrix requant of orcarouter's FP8 abliteration.