mradermacher/Qwen3-VL-2B-Instruct-abliterated-GGUF
Abliterated (fewer layers) vs prithivMLmods/Qwen3-VL-2B-Instruct-abliterated-v1. mradermacher/Qwen3-VL-2B-Instruct-abliterated-GGUF matches part of prithivMLmods/Qwen3-VL-2B-Instruct-abliterated-v1 (310 of its 625 tensors; the rest of prithivMLmods/Qwen3-VL-2B-Instruct-abliterated-v1 has no counterpart). On the shared part: mradermacher/Qwen3-VL-2B-Instruct-abliterated-GGUF is a GGUF (F16) of different weights. Looks like abliteration (refusal removal): only the output projections (down_proj, o_proj) changed, in 28 of 28 layers. Abliteration edits exactly these matrices to remove a "refusal direction"; everything else is untouched.
Tensors changed: Embeddings 0/1, Norms 0/113, MLP 28/84, Attention 28/112