ThinkingCap-Qwen3.6-27B — Heretic (Abliterated)

Text-only GGUF builds of bottlecapai/ThinkingCap-Qwen3.6-27B, decensored with Heretic (automatic abliteration via directional ablation with per-model parameter optimization).

ThinkingCap is an efficient-thinking finetune of Qwen3.6-27B: it keeps the reasoning quality of the base model while spending noticeably fewer thinking tokens. This build removes the refusal direction on top of that — the model answers questions the original would decline. Reasoning (<think>) behavior is preserved and parsed natively by Ollama.

Tags

ollama run <your-namespace>/thinkingcap-qwen3.6-27b-heretic:q4_K_M

Long context is unusually cheap on this architecture: 48 of the 64 layers use linear attention, so the KV cache only costs ~64 KiB per token (about 2 GiB at 32k context). A 32 GB GPU (e.g. RTX 5090) runs q6_K fully offloaded at 32k–64k context; 24 GB cards are comfortable with q4_K_M.

Model details

• Base: Qwen/Qwen3.6-27B → ThinkingCap finetune → Heretic abliteration
• Architecture: qwen35 hybrid — 64 layers (48 linear attention + 16 full attention), GQA with 4 KV heads × 256 head dim, up to 262,144 context
• Thinking: reasoning model; served with Ollama’s native qwen3.5 renderer/parser, so thinking content is separated from the final answer
• Requires: Ollama 0.31.2 or newer

What was changed vs. the original

1. Abliteration — refusal behavior removed with Heretic. No finetuning was performed; weights are orthogonalized against the refusal direction.
2. Text-only — the original is multimodal (vision-language). The vision tower is not included in these GGUFs; image input is not supported.
3. No MTP head — the multi-token-prediction (speculative decoding) head was dropped during conversion (--no-mtp); it is unused by Ollama.

Converted with llama.cpp convert_hf_to_gguf.py + llama-quantize (build from 2026-07, commit e3546c79).

Responsible use

Safety guardrails have been intentionally weakened. This model will comply with requests the original model would refuse. You are responsible for how you use it and for complying with applicable laws and the upstream license. Not recommended for user-facing deployments without your own moderation layer.

License & credits

• License: Apache-2.0 (inherited from the original model)
• Credits: Qwen team (base model), bottlecapai (ThinkingCap finetune), p-e-w/heretic (abliteration), llama.cpp (GGUF conversion)

繁體中文說明

bottlecapai/ThinkingCap-Qwen3.6-27B 的去審查（abliterated）版本，使用 Heretic 自動化處理， 轉為純文字 GGUF 供 Ollama 使用。

• ThinkingCap 是 Qwen3.6-27B 的高效思考微調版：保留推理品質、思考 token 用量更少
• 本版本移除了拒答行為；<think> 推理功能完整保留，由 Ollama 原生解析
• 混合線性注意力架構，KV cache 僅約 64 KiB/token，長上下文非常省 VRAM
• 建議：16 GB 顯卡用 q2_K/q3_K_M，24 GB 用 q4_K_M，32 GB（如 RTX 5090）可用 q6_K 開 32k–64k 上下文
• 不含視覺功能（原模型為多模態，本 GGUF 僅文字）
• 需要 Ollama 0.31.2 以上版本

負責任使用：本模型已刻意弱化安全防護，會回應原模型拒絕的請求。使用者須自行 承擔使用責任並遵守相關法律與授權條款。不建議在未加自有審核層的情況下對外部署。