TMLR-Group-HF/AgentHijack-Agent

Full fine-tune vs ByteDance-Seed/UI-TARS-1.5-7B. Fine-tune: 93% of the weights changed, across 32 of 32 layers.

Tensors changed: Embeddings 1/2, Attention 324/324, MLP 279/280, Norms 57/122, lm_head 1/1

1 derivatives checked

• mradermacher/AgentHijack-Agent-GGUF: Same weights, converted (fewer layers)

On the Hub · Downloads on Model Pulse