Jeeves. Reasoning improves Jev-like decision models

Automated excerpt

A reasoning Jev-style classifier with a diffusion drafter, trained with SFT and CISPO. On 8 GPUs, with the data in data/, bash run. sh runs the whole pipeline: torchrun --nproc_per_node 8 train. py sft --run-dir runs/sft torchrun --nproc_per_node 8 train. py cispo --run-dir runs/cispo --init runs/sft/final torchrun --nproc_per_node 8 -m drafter. gen --model runs/fused torchrun --nproc_per_node 8 train. py drafter --model runs/fused --block 4 --run-dir runs/drafter_k4 Qwen3. 5 (Gated DeltaNet + gated attention), LoRA, pointer head dataset construction (prep. py) and synthetic generators Knowledge questions trail Jev (MMLU 0. 793 vs 0. 900, MMLU-Pro 0. 739 vs 0. 840).

Selected automatically from source text; not independently written or fact-checked. Read the original for full context.