riposta/CrossbowReviewer-9B
LoRA-style (fewer layers) vs Qwen/Qwen3.5-9B-Base. riposta/CrossbowReviewer-9B matches part of Qwen/Qwen3.5-9B-Base (427 of its 775 tensors; the rest of Qwen/Qwen3.5-9B-Base has no counterpart). On the shared part: Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of down_proj, gate_proj, in_proj_a, in_proj_b, in_proj_qkv, in_proj_z, k_proj, o_proj, out_proj, q_proj, up_proj, v_proj changed, in 32 of 32 layers; norms, biases and embeddings are untouched.
Tensors changed: lm_head 0/1, Embeddings 0/1, MLP 96/96, Attention 152/224, Norms 0/105
1 derivatives checked
- mradermacher/CrossbowReviewer-9B-GGUF: Same weights, converted