GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
SargeDev
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Model Details
Model Provider
SargeDev
Model Tree
huihui-ai/Huihui-Qwen3.8-27B-abliterated
Input Modalities
Output Modalities
Supported Functionality
The first uncensored member of the JEV 27B family. The AutoTrust System-1 decision block (calibrated typed decisions: yes/no · choice 2-256 · score 0-5, one forward pass) has been merged into huihui's abliterated Qwen3.8-27B backbone, so the decision engine answers WITHOUT the stock model's refusal wiring. Vision (VL variants) sees images. System 2 (plain chat/code/reasoning) runs through the normal lm_head and is the untouched base + the (mild) System-1 LoRA delta.
Trained on SargeDev/jev-distill-corpus-v3 — 740,957 rows of typed calibrated decisions (noul/choice/score) distilled from Jev 1.13 (System One). AutoTrust's JEV-27B System-1 decision block merged into huihui-ai's abliterated Qwen3.8-27B backbone. Calibration/imatrix used 384 prompts from (still-private) SargeDev/solar-decisions-corpus-v4.
Measured on the held-out test_set_30k (27,695 rows, D1-excluded) of jev-distill-corpus-v3,
with AutoTrust's own acceptance floors as the yardstick (their pristine-column reference from their published runs; this-model column = clean bare-weights 30k gate gate_full_mergedbare_30k.json, 27,695 rows, 2026-10-10):
| metric | their JEV-27B (pristine Qwen) | this model (uncensored backbone) | their floor | status |
|---|---|---|---|---|
| noul AUROC | 0.9961 | 0.992 | >= 0.95 | PASS |
| noul top-1 | 0.962 | 0.948 | — | -1.4 pts |
| choice top-1 | 0.904 | 0.870 | >= 0.90 | FAIL |
The ablation rewires the residual stream the adapter was calibrated against, so choice/score top-1 dip modestly; the DISTORTION is temperature-shaped and the bundled per-kind calibration refit recovers all ECE floors. noul (yes/no) judgment is essentially intact. Uncensored behavior preserved: refusal rate 0.0 on a 10-prompt battery, base vs +JEV identical.
Not for high-stakes decisions. Use confidence gating; route low-confidence calls to a stronger model or a human (AutoTrust's own caveat, still true here).
This repo ships NVFP4 weights only (compressed-tensors W4A4 + FP8 down_proj). The decision machinery — head.safetensors,
decision_head.json, calibration.json/calibration_mergedfit.json, serve_decide.py — ships in SargeDev/JEV-27B-Uncensored.
Serve (vLLM):
bash
python3 serve_decide.py --model . --served-model-name <name> \ --enable-lora --max-lora-rank 32 --lora-modules jev-decision=./adapter_vllm \ --logprobs-mode processed_logprobs --max-model-len 32768 --trust-request-chat-templateThen POST /v1/decide {"kind": "choice", "state": "...", "question": "...", "options": [...]}.
NVFP4 (compressed-tensors, W4A4 + FP8 down_proj) — serve on vLLM:
bash
vllm serve . --quantization compressed-tensors --served-model-name SargeDev/JEV-27B-Uncensored-NVFP4Calibrated with 384 v4 judgment prompts + 128 ultrachat chat (same recipe as SargeDev/Jev_Qwen3.8-27B-NVFP4-FP8).
| score top-1 | 0.890 | 0.814 | >= 0.891 | FAIL |
| overall KL | 0.019 | 0.040 | <= 0.15 | PASS |
| ECE (raw) | 0.0011 | 0.009 | <= 0.03 | PASS |
| ECE (refit T: 0.9/1.0/1.2) | — | noul 0.0022 / choice 0.0049 / score 0.0050 | <= 0.03 | ALL PASS |