# Gestalt Labs: jeff

> A rank-16 LoRA on Qwen3-4B-Instruct-2507 that answers choice, noul and score questions over labels named at call time, reading probabilities from the model's own token distribution rather than generated text. Tuned for checking a claim against supplied evidence.

- Page: https://systemonemodels.ai/gestalt-labs/jeff
- API: https://api.systemonemodels.ai/v1/models/gestalt-labs/jeff
- Download: `pip install systemonemodels && systemone pull gestalt-labs/jeff`

## Facts

| | |
|---|---|
| Maker | Gestalt Labs (https://systemonemodels.ai/gestalt-labs) |
| Decides | choice, score, noul |
| Architecture | jeff |
| Base model | qwen/qwen3-4b-instruct-2507 |
| Parameters | 4.0B |
| Licence | apache-2.0 |
| Availability | Open weights |
| Released | 2026-09-19 |
| Latest version | 2026.09 |

## Reported evaluation

Suite: Gestalt Labs fact-check audit (9,730 human-labelled rows; also used for error analysis), the maker's own run. Numbers are the publisher's own.

- Decision accuracy: 81.8%
- Calibration error (ECE): 0.081

## Model card

<!-- generated by scripts/seed_catalog.py; edit content/models/catalog.yaml -->

# Jeff 1

A rank-16 LoRA on Qwen3-4B-Instruct-2507 that answers choice, noul and score questions over labels named at call time, reading probabilities from the model's own token distribution rather than generated text. Tuned for checking a claim against supplied evidence.

A PEFT adapter (11.8M trainable parameters), so the base model is needed. Choice uses first-token scoring when options start with different tokens; Score uses whole-sequence scoring, one extra pass per level. Trained on 12,119 rows for fact-checking against supplied evidence; it does not retrieve sources. On Gestalt Labs' own audit (9,730 human-labelled fact-check rows from VitaminC, FEVER, SciFact and Climate-FEVER, also used for error analysis) it scores 0.818 accuracy and ECE 0.081; the maker's own re-measurement of Jev 1.13.0 on the same rows gives 0.828 and 0.093. Weak on not-enough-info cases, and it marks refuted claims as supported more often. English only; latency not measured. The adapter is Apache-2.0, but about 7,000 training labels came from Jev 1.13.0 and 413 rows from SciFact (CC BY-NC 2.0); check both before commercial use. Ships a train-your-own guide and a local /v1/systemone server. Not affiliated with TypeSafe AI.

## What it decides

- **choice** — picks one option from a set
- **score** — places the input on an ordered scale
- **noul** — answers a yes/no question with one probability

## At a glance

| | |
|---|---|
| Parameters | 4B |
| Base model | `Qwen/Qwen3-4B-Instruct-2507` |
| Maker | Gestalt Labs |
| Released | 2026-09-19 |
| License | apache-2.0 |
| Reported accuracy | 81.8% |

## Get the weights

```bash
pip install systemonemodels
systemone pull gestalt-labs/jeff
```

The files are served from the maker's Hugging Face repository, [`GestaltLabs/Jeff-1`](https://huggingface.co/GestaltLabs/Jeff-1), and verified against the checksums recorded here.

## Read more

- [Model card](https://huggingface.co/GestaltLabs/Jeff-1)
- [Code and training guide](https://github.com/Gestalt-Lab/jeff)

---

*This page was opened by System One for Gestalt Labs, who can claim the organisation and take it over at any time.*

---

From System One Models — https://systemonemodels.ai/ · every System One model: https://systemonemodels.ai/system-one-models