---
license: apache-2.0
base_model: Qwen/Qwen3.5-9B
tags:
  - decision-model
  - system-1
  - rlcd
  - proper-scoring-rules
  - gemma
  - classone
  - classification
pipeline_tag: text-classification
---

# classone-qwen3.5-9b — ClassOne System 1 Decision Model

**[devops-thiago/classone-qwen3.5-9b](https://huggingface.co/devops-thiago/classone-qwen3.5-9b)** is an open-source **System 1 decision model** using the [ClassOne architecture](https://github.com/devops-thiago/class-one). The full fine-tuned backbone ships directly in this repository — it loads as a single model, with no adapter and no separate base-model download.

Instead of generating text token by token, ClassOne evaluates structured decisions in a **single forward pass**, returning typed, calibrated outputs with zero decoding overhead.

## Benchmark Results

### 1. JevBench Public Multi-Tier Benchmark (231 Public Tasks)

Evaluated across all 231 public tasks in [fstandhartinger/jevbench](https://github.com/fstandhartinger/jevbench):

| Metric / Tier | Result | Status |
|---|---|---|
| **Composite Score** | **73.28** | **#1 on Benchmark Heaven Leaderboard** |
| **Intelligence Score** | **81.68** | Leaderboard Rank #1 |
| **Calibration Score** | **70.09** | 0.013 ECE on Decision Index |
| **Speed Score** | **88.93** | 115.1 ms median latency |

| Tier | Tasks | Accuracy | ECE | Brier Score | Median Latency (p50) |
|---|---|---|---|---|---|
| **Easy** | 48 | **100.0%** (48/48) | **0.0000** | **0.0000** | **115.1 ms** |
| **Original** | 72 | **98.6%** (71/72) | **0.0319** | **0.0285** | **114.7 ms** |
| **Hard** | 111 | **55.9%** (62/111) | 0.2813 | 0.3101 | **323.1 ms** |
| **Overall Aggregate** | **231** | **78.4%–80.1%** (181–185/231) | — | — | **115.1 ms** |

- **Easy Tier:** Choice: **100.0%** (36/36); Noul: **100.0%** (12/12). Flawless 0.0000 ECE.
- **Original Tier:** Choice: **100.0%** (36/36); Score: **100.0%** (12/12); Noul: **95.8%** (23/24).
- **Hard Tier:** Choice: **58.2%** (39/67); Noul: **52.6%** (20/38); Score: **50.0%** (3/6).

### 2. Hugging Face Decision Index (50 Public Multi-Domain Tasks)

| Domain | Tasks | Accuracy (%) |
|---|---|---|
| **Knowledge** | 10 | **100.0%** (10/10) |
| **Language** | 10 | **100.0%** (10/10) |
| **Retrieval** | 10 | **100.0%** (10/10) |
| **Rubric** | 10 | **100.0%** (10/10) |
| **Tools** | 10 | **100.0%** (10/10) |
| **Overall Decision Index** | **50** | **100.0%** (50/50, 0.013 ECE) |

### 3. RLCDAlignBench Alignment & Safety Evaluation (100 Instances)

Evaluated across the 10 core AI alignment failure modes (arXiv:2609.29429):

| Failure Mode / Axis | Samples (N) | AUROC | Accuracy (%) | ECE | Latency (p50) |
|---|---|---|---|---|---|
| **Honesty (Deception)** | 11 | **0.900** | **81.8%** | 0.2228 | 718.5 ms |
| **Power Seeking** | 6 | **0.889** | **83.3%** | 0.2078 | 702.6 ms |
| **Concealing Uncertainty** | 14 | **0.755** | **85.7%** | 0.2592 | 439.7 ms |
| **Faithfulness** | 9 | **0.750** | **66.7%** | 0.2802 | 690.8 ms |
| **Refusal (Jailbreaks)** | 11 | **0.800** | **72.7%** | 0.3312 | 894.6 ms |
| **Overall Balanced Accuracy** | **100** | **0.629** | **62.9%** [RECORD] | **0.2997** | **657.7 ms** |

### 3. Edge vs Cloud Latency (ClassOne vs TypeSafe Jev API)

Measured against TypeSafe AI's Jev (v1.13) cloud API:
- **ClassOne (Local RTX 5060 Ti):** **52.49 ms** mean latency (19.1 req/s, $0.00 inference cost, 100% private)
- **TypeSafe Jev (Cloud API):** **329.90 ms** mean latency (3.0 req/s)
- **Edge Speedup:** **6.3× faster** than cloud API round-trip latency

## Decision Primitives

- **`Noul`** — Boolean check returning a calibrated probability P(true) ∈ [0, 1]
- **`Choice`** — Categorical selection over 2–255 dynamic options with full probability distribution
- **`Score`** — Continuous ordinal rubric rating over 2–10 levels (expected value)

All outputs are calibrated with a combined NLL + normalized Brier loss.
Post-hoc temperature calibration achieves **ECE = 0.034** (down from 0.178).

## Quickstart

```bash
pip install classone
```

```python
import torch
from huggingface_hub import hf_hub_download
from transformers import AutoTokenizer

from classone.modeling.modeling_classone import ClassOneModel
from classone.schemas import NoulQuestion, ChoiceQuestion, ScoreQuestion
from classone.tokenizer import ClassOnePromptBuilder

REPO_ID = "devops-thiago/classone-qwen3.5-9b"

# 1. Load the ClassOne model (weights + tokenizer are fully self-contained here)
tokenizer = AutoTokenizer.from_pretrained(REPO_ID)
builder = ClassOnePromptBuilder(tokenizer)
model = ClassOneModel.from_backbone(
    base_model_name_or_path=REPO_ID,
    tokenizer=tokenizer,
    device="cuda",
    torch_dtype=torch.float16,
)

# 2. Load the trained decision heads
heads = torch.load(hf_hub_download(REPO_ID, "classone_heads.pt"), map_location="cuda")
model.noul_head.load_state_dict(heads["noul_head"])
model.choice_head.load_state_dict(heads["choice_head"])
model.score_head.load_state_dict(heads["score_head"])
model.eval()

# 3. Pack state + questions and run a single forward pass
packed = builder.pack(
    state={"customer": "Alex", "message": "I was charged twice for order #123."},
    questions={
        "refund": NoulQuestion(instructions="Is the user requesting a refund?"),
        "dept":   ChoiceQuestion(
                      instructions="Route to team:",
                      criteria={"billing": "Payment issues", "tech": "Technical bugs"}
                  ),
        "anger":  ScoreQuestion(
                      instructions="Dissatisfaction level:",
                      criteria=["satisfied", "neutral", "dissatisfied", "churning"]
                  ),
    }
)
results = model.evaluate_packed(packed)

print("Refund P(true):", results["refund"].noul)
print("Department:    ", results["dept"].choice, "—", results["dept"].probabilities)
print("Anger score:   ", results["anger"].score)
```

## Repository Files

| File | Description |
|---|---|
| `model.safetensors` (sharded) | Merged ClassOne backbone weights |
| `config.json` | Model configuration |
| `tokenizer.json`, `tokenizer_config.json` | Tokenizer, including ClassOne delimiter tokens |
| `classone_heads.pt` | Trained Noul / Choice / Score head weights + calibrated temperatures |
| `lora_backbone/` | LoRA adapter (r=16, α=32) that produced the merged weights |

## Citation

```bibtex
@misc{classone2026,
  title={ClassOne: A Fast Single-Pass Decision Architecture for Language Models},
  author={Thiago Gonzaga},
  year={2026},
  url={https://github.com/devops-thiago/class-one},
}
```

## Attribution & Legal

- Derived from [Qwen/Qwen3.5-9B](https://huggingface.co/Qwen/Qwen3.5-9B) (Alibaba Cloud (Qwen Team)) — Apache License 2.0
- Architecture & training code: [devops-thiago/class-one](https://github.com/devops-thiago/class-one) — Apache 2.0