---
license: apache-2.0
library_name: kev.js
pipeline_tag: text-classification
# Immediate parents: Jared's Hub checkpoints. Their cards already name the Qwen3.5 bases, so the model tree is
# Qwen → kev-* → this repo (listing Qwen here would flatten that hop).
base_model:
- jaredpalmer/kev-0.8b
- jaredpalmer/kev-4b
- jaredpalmer/kev-9b
tags: [kev, decision-model, onnx, onnxruntime-web, webgpu, quantized, int8]
---

# kev.js weights

Browser-ready exports of [Kev](https://github.com/jaredpalmer/kev), Jared Palmer's family of small decision models.
They answer yes/no, multiple-choice and rating questions with calibrated probabilities, and run in a browser on
WebGPU through [@ai-ecoverse/kev.js](https://github.com/ai-ecoverse/kev.js). This repo holds only converted
weights: no training, evaluation or model design here is ours.

```js
import * as ort from "onnxruntime-web/webgpu";
import { loadKev } from "@ai-ecoverse/kev.js";

const kev = await loadKev("https://huggingface.co/ai-ecoverse/kev.js/resolve/main/kev-0.8b", { ort, variant: "q8f32" });
const res = await kev.systemOne({
  state: "I was charged twice. Please fix this ASAP.",
  questions: { billing: { type: "noul", instructions: "Is this ticket about billing?" } },
});
```

## Contents

| Folder | Variant | Download | Base | Source checkpoint |
|---|---|---|---|---|
| `kev-0.8b` | `q8f32` | 0.84 GB | `Qwen/Qwen3.5-0.8B-Base` | `jaredpalmer/kev-0.8b@9a45d25eb2ab761841196625383fa1dff0e56c1e` |
| `kev-0.8b` | `q8` | 0.80 GB | `Qwen/Qwen3.5-0.8B-Base` | `jaredpalmer/kev-0.8b@9a45d25eb2ab761841196625383fa1dff0e56c1e` |
| `kev-4b` | `q8f32` | 4.69 GB | `Qwen/Qwen3.5-4B-Base` | `jaredpalmer/kev-4b@139fdd94f1b6a6ad80cc15e08fcb99cac885a101` |
| `kev-9b` | `q8f32` | 8.84 GB | `Qwen/Qwen3.5-9B-Base` | `jaredpalmer/kev-9b@2629c06a5aeb0feb3b9783bafed17ed8f39ecf5c` |

Every bundle is int8 weights with fp32 activations, split into 32 MB files so any CDN or proxy can serve them.
`manifest.json` lists the files, their sizes, the tokenizer and the pointer head, plus the measured deviation from
the original fp32 PyTorch model on a fixture set.

## Provenance and licenses

- Models and training: [jaredpalmer/kev](https://github.com/jaredpalmer/kev) (Apache-2.0). Source checkpoints:
  `jaredpalmer/kev-0.8b@9a45d25eb2ab761841196625383fa1dff0e56c1e`, `jaredpalmer/kev-4b@139fdd94f1b6a6ad80cc15e08fcb99cac885a101`, `jaredpalmer/kev-9b@2629c06a5aeb0feb3b9783bafed17ed8f39ecf5c`.
- Base models: [Qwen3.5](https://huggingface.co/Qwen) (Apache-2.0) — via the Kev checkpoints above (`Qwen/Qwen3.5-0.8B-Base`, `Qwen/Qwen3.5-4B-Base`, `Qwen/Qwen3.5-9B-Base`).
- Architecture described in [Jev's Architecture Unmasked](https://archerhume.com/posts/jevs-architecture-unmasked).
  The API shapes follow [TypeSafe's System One](https://docs.typesafe.ai/api); Jev is TypeSafe's hosted model and is
  not affiliated with this repo.
- Conversion: LoRA merged in fp32, exported with the onnxruntime-genai model builder without the LM head, embeddings
  quantized to int8 per row. Details in the [kev.js README](https://github.com/ai-ecoverse/kev.js#readme).
- `-vision` folders: the same decoder with an `image_embeds` input, plus the base model's own Qwen3.5 vision tower and
  patch merger (fp16 weights), unmodified and not trained with Kev. Accuracy on images is in the kev.js README.