Quantized builds of Inkling 2 Mini (Thinking Machines Lab): GGUF Q4_K_M, GGUF IQ3_XXS and MLX 4-bit.
AI & ML interests
Interpretability for open models. We read models from the inside, neuron by neuron. Then we make them better.
Recent Activity
Built on Arete-27B (Nous Research), with the remaining CCP censorship removed. BF16, GGUF and MLX.
-
InterpreticAI/Qwen3.8-27B-Uncensored-v3
Image-Text-to-Text • 27B • Updated • 367 • 9 -
InterpreticAI/Qwen3.8-27B-Uncensored-v3-GGUF
Image-Text-to-Text • 27B • Updated • 6.93k • 1 -
InterpreticAI/Qwen3.8-27B-Uncensored-v3-IQ3_XXS-GGUF
Image-Text-to-Text • 27B • Updated • 1.44k -
InterpreticAI/Qwen3.8-27B-Uncensored-v3-MLX-4bit
Image-Text-to-Text • 27B • Updated
Qwen3.8-27B with political censorship removed. BF16, FP8 and GGUF, with MTP.
-
InterpreticAI/Qwen3.8-27B-Uncensored-v2
Image-Text-to-Text • 28B • Updated • 366 • 4 -
InterpreticAI/Qwen3.8-27B-Uncensored-v2-FP8
Image-Text-to-Text • 28B • Updated • 265 -
InterpreticAI/Qwen3.8-27B-Uncensored-v2-GGUF
Image-Text-to-Text • 27B • Updated • 10.1k • 2 -
InterpreticAI/Qwen3.8-27B-Uncensored-v2-IQ3_XXS-GGUF
Image-Text-to-Text • 27B • Updated • 2.5k
Qwen4-27B with political censorship removed. Weights land here as soon as Qwen4-27B is released.
Uncensored model: Qwen3.5-4B with political censorship removed. BF16, GGUF, browser build, and Qwenputer (RISC-V computer in the weights).
-
InterpreticAI/Qwen3.5-4B-Uncensored-v2
Image-Text-to-Text • 5B • Updated • 139 -
InterpreticAI/Qwen3.5-4B-Uncensored-v2-GGUF
Image-Text-to-Text • 4B • Updated • 2.96k • 1 -
InterpreticAI/Qwen3.5-4B-Uncensored-v2-Browser-GGUF
Text Generation • 4B • Updated • 721 -
InterpreticAI/Qwenputer-4B-Uncensored-v2
Text Generation • 9B • Updated • 423 • 1
Quantized builds of Inkling 2 Mini (Thinking Machines Lab): GGUF Q4_K_M, GGUF IQ3_XXS and MLX 4-bit.
Qwen4-27B with political censorship removed. Weights land here as soon as Qwen4-27B is released.
Built on Arete-27B (Nous Research), with the remaining CCP censorship removed. BF16, GGUF and MLX.
-
InterpreticAI/Qwen3.8-27B-Uncensored-v3
Image-Text-to-Text • 27B • Updated • 367 • 9 -
InterpreticAI/Qwen3.8-27B-Uncensored-v3-GGUF
Image-Text-to-Text • 27B • Updated • 6.93k • 1 -
InterpreticAI/Qwen3.8-27B-Uncensored-v3-IQ3_XXS-GGUF
Image-Text-to-Text • 27B • Updated • 1.44k -
InterpreticAI/Qwen3.8-27B-Uncensored-v3-MLX-4bit
Image-Text-to-Text • 27B • Updated
Uncensored model: Qwen3.5-4B with political censorship removed. BF16, GGUF, browser build, and Qwenputer (RISC-V computer in the weights).
-
InterpreticAI/Qwen3.5-4B-Uncensored-v2
Image-Text-to-Text • 5B • Updated • 139 -
InterpreticAI/Qwen3.5-4B-Uncensored-v2-GGUF
Image-Text-to-Text • 4B • Updated • 2.96k • 1 -
InterpreticAI/Qwen3.5-4B-Uncensored-v2-Browser-GGUF
Text Generation • 4B • Updated • 721 -
InterpreticAI/Qwenputer-4B-Uncensored-v2
Text Generation • 9B • Updated • 423 • 1
Qwen3.8-27B with political censorship removed. BF16, FP8 and GGUF, with MTP.
-
InterpreticAI/Qwen3.8-27B-Uncensored-v2
Image-Text-to-Text • 28B • Updated • 366 • 4 -
InterpreticAI/Qwen3.8-27B-Uncensored-v2-FP8
Image-Text-to-Text • 28B • Updated • 265 -
InterpreticAI/Qwen3.8-27B-Uncensored-v2-GGUF
Image-Text-to-Text • 27B • Updated • 10.1k • 2 -
InterpreticAI/Qwen3.8-27B-Uncensored-v2-IQ3_XXS-GGUF
Image-Text-to-Text • 27B • Updated • 2.5k