Uniphore/actio-ui-7b-rlvr
Partial fine-tune vs Uniphore/actio-ui-7b-sft. Partial change: 87% of the weights changed (attention, embeddings, mlp, norms), in 32 of 32 layers.
Tensors changed: Embeddings 1/2, Attention 319/324, MLP 273/280, Norms 57/122, lm_head 0/1
3 derivatives checked
- mradermacher/actio-ui-7b-rlvr-i1-GGUF: Quantization (fewer layers)
- mradermacher/actio-ui-7b-rlvr-GGUF: Same weights, converted (fewer layers)
- prithivMLmods/actio-ui-7b-rlvr-GGUF: Same weights, converted (fewer layers)