Uncensored Nemotron 3.5 Bolt, abliterated inside NVIDIA's shipped 2-bit GGUF.
Josh Bolding PRO
BoldingBuilds
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
new activity 2 days ago
BoldingBuilds/AREX-2-Abliterated-GGUF:Best for me! updated a model 8 days ago
BoldingBuilds/Ternary-Bonsai-2-27B-Abliterated-v2-PQ2_0-MTP-GGUF updated a model 8 days ago
BoldingBuilds/Ternary-Bonsai-2-27B-Abliterated-v2-PTQ1_0-GGUFOrganizations
The </think> row edit that stops thinking-mode runaway. ThinkFix = abliterated + fix; Reasoning-Termination-Fix = stock weights + fix only.
-
BoldingBuilds/Qwen3.8-27B-Abliterated-ThinkFix-GGUF
Text Generation • 27B • Updated • 1.89k • 5 -
BoldingBuilds/Qwen3.8-27B-Reasoning-Termination-Fix-GGUF
Text Generation • 27B • Updated • 468 • 1 -
BoldingBuilds/Qwen3.8-Flash-Next-Reasoning-Termination-Fix-GGUF
Text Generation • 177B • Updated • 128 • 2
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.
BAAI's AREX-2 agent fine tune of Qwen3.8 with refusals removed, GGUF quants.
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 • 45.5k • 46 -
BoldingBuilds/Ternary-Bonsai-2-27B-Abliterated-v2-PTQ1_0-GGUF
Text Generation • 27B • Updated • 9.41k • 3 -
BoldingBuilds/Ternary-Bonsai-2-27B-Abliterated-PQ2_0-MTP-GGUF
Text Generation • 27B • Updated • 6.25k • 31 -
BoldingBuilds/Ternary-Bonsai-2-27B-Abliterated-PTQ1_0-GGUF
Text Generation • 27B • Updated • 12.4k • 84
ShimQuant GGUF builds of NVIDIA Nemotron 3.5 Lightning 30B-A3B, stock and abliterated.
Uncensored Nemotron 3.5 Bolt, abliterated inside NVIDIA's shipped 2-bit GGUF.
BAAI's AREX-2 agent fine tune of Qwen3.8 with refusals removed, GGUF quants.
The </think> row edit that stops thinking-mode runaway. ThinkFix = abliterated + fix; Reasoning-Termination-Fix = stock weights + fix only.
-
BoldingBuilds/Qwen3.8-27B-Abliterated-ThinkFix-GGUF
Text Generation • 27B • Updated • 1.89k • 5 -
BoldingBuilds/Qwen3.8-27B-Reasoning-Termination-Fix-GGUF
Text Generation • 27B • Updated • 468 • 1 -
BoldingBuilds/Qwen3.8-Flash-Next-Reasoning-Termination-Fix-GGUF
Text Generation • 177B • Updated • 128 • 2
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 • 45.5k • 46 -
BoldingBuilds/Ternary-Bonsai-2-27B-Abliterated-v2-PTQ1_0-GGUF
Text Generation • 27B • Updated • 9.41k • 3 -
BoldingBuilds/Ternary-Bonsai-2-27B-Abliterated-PQ2_0-MTP-GGUF
Text Generation • 27B • Updated • 6.25k • 31 -
BoldingBuilds/Ternary-Bonsai-2-27B-Abliterated-PTQ1_0-GGUF
Text Generation • 27B • Updated • 12.4k • 84
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.