mradermacher/Qwen3-VL-4B-Thinking-abliterated-GGUF
Abliterated (fewer layers) vs prithivMLmods/Qwen3-VL-4B-Thinking-abliterated-v1. mradermacher/Qwen3-VL-4B-Thinking-abliterated-GGUF matches part of prithivMLmods/Qwen3-VL-4B-Thinking-abliterated-v1 (398 of its 713 tensors; the rest of prithivMLmods/Qwen3-VL-4B-Thinking-abliterated-v1 has no counterpart). On the shared part: mradermacher/Qwen3-VL-4B-Thinking-abliterated-GGUF is a GGUF (F16) of different weights. Looks like abliteration (refusal removal): only the output projections (down_proj, embed_tokens, o_proj) changed, in 36 of 36 layers. Abliteration edits exactly these matrices to remove a "refusal direction"; everything else is untouched.
Tensors changed: Embeddings 1/1, Norms 0/145, MLP 36/108, Attention 36/144