sn92-gen-v52a: SN92 generalist decision model

A LoRA fine-tune (rank 16, merged into the weights) of danyloooah/sn92-generalist-m1-core2 (revision c78c0a4a), itself a weight-space merge of fine-tunes of baselabs/mammoth-1 (revision 03007ab3c4481bf01c71ef63adbbbc4b2093d71a, MIT, Base Labs / Baseten; post-trained from Qwen3.5-27B, Apache-2.0). REAX joint-readout layout (Qwen3_5Model backbone, answer-code readout), unchanged from the parent.

Training data: REAX eval-v5.1-style decision rows plus single-choice and yes/no rows from the train split of tasksource/tasksource-jev-typed-decisions, with a KL term to the parent.

The parents' learned changes are their authors' work and the credit for them is theirs; see LICENSE and NOTICE. Not affiliated with, endorsed by or sponsored by Base Labs / Baseten, the Qwen team, REAX or the parents' authors.