mradermacher/AdaReasoner-7B-GGUF
Full fine-tune (fewer layers) vs AdaReasoner/AdaReasoner-7B-Randomized. mradermacher/AdaReasoner-7B-GGUF matches part of AdaReasoner/AdaReasoner-7B-Randomized (339 of its 729 tensors; the rest of AdaReasoner/AdaReasoner-7B-Randomized has no counterpart). On the shared part: mradermacher/AdaReasoner-7B-GGUF is a GGUF (F16) of different weights. Fine-tune: 93% of the weights changed, across 28 of 28 layers.
Tensors changed: Embeddings 0/1, MLP 84/84, Norms 53/57, Attention 196/196, lm_head 1/1