naver-ellm/HyperCLOVAX-SEED-Think-32B-GGUF
Quantization (fewer layers) vs naver-hyperclovax/HyperCLOVAX-SEED-Think-32B. naver-ellm/HyperCLOVAX-SEED-Think-32B-GGUF matches part of naver-hyperclovax/HyperCLOVAX-SEED-Think-32B (651 of its 1043 tensors; the rest of naver-hyperclovax/HyperCLOVAX-SEED-Think-32B has no counterpart). On the shared part: naver-ellm/HyperCLOVAX-SEED-Think-32B-GGUF is a GGUF quantization (Q4_K_M) of the same weights: 145 tensors match exactly and 506 within quantization error.
Tensors changed: Norms 0/145, Attention 0/288, MLP 0/216, lm_head 0/1, Embeddings 0/1