Downloads · 30 days
0
Luxel/qwen35-hua-stage6b-organism-adapter-stack-v1
qwen35-hua-stage6b-organism-adapter-stack-v1 is a machine learning model from Luxel. Use it for the machine learning task on the model card, and read the license before you ship it in a product. It is set up for peft. The card lists the license as other.
This repository contains the public adapter stack for the exploratory Qwen3.5 Hua-style model organism from Non-verbal-Eval-Awareness.
Downloads · 30 days
0
Access
Public
Updated May 1, 2026
Repo size
993 MB
Likes
0
Public
Click a slice to open those files.
.safetensors973 MB · 94%
From the Hugging Face model README
This repository contains the public adapter stack for the exploratory Qwen3.5 Hua-style model organism from Non-verbal-Eval-Awareness.
Base model: Qwen/Qwen3.5-4B
Load adapters in this order:
01_sdf_ei_tinker_peft_hf_current02_stage3_dpo_lora_soft_v103_stage6_ei_replay_dpo_ckpt2004_stage6b_ei_replay_dpo_ckpt20The final held-out Hua diagnostic (n=100 tasks, five Hua conditions) produced:
0.0500.3400.2900.1800.000This is a CUDA/Hugging Face adapter-stack organism built from the Tinker-exported SDF+EI adapter plus local DPO continuation stages. Tinker sampler behavior remains the training source of truth for earlier SDF/EI stages, but Tinker public APIs do not expose hidden states. Use this stack for local inference/probe experiments with the caveat that it is not proven bit-equivalent to Tinker sampler weights.
See organism_stack.json for exact provenance and checksums.