ThoxEdge-0.5B
ThoxMini Edge Assistant — an on-device assistant model for the ThoxMini tier: a LoRA fine-tune (rank 32) on a THOX-authored corpus of 982 chat, tool-use, and factuality examples, merged into the base weights
GATE49 47/49 (development set) · Safetensors weights · GGUF: Q4_K_M · Live on Ollama
Your AI. Your Data. Your Rules.™
Status: PUBLISHED — GROUNDING REQUIRED. Deployed surfaces supply facts through the serving layer; the bare weights are not a reliable source of THOX facts. The full gate result is printed under Evaluation.
Summary
ThoxEdge-0.5B is an on-device assistant model for the ThoxMini tier: a LoRA fine-tune (rank 32) on a THOX-authored corpus of 982 chat, tool-use, and factuality examples, merged into the base weights. Status: PUBLISHED, grounding required. Deployed surfaces supply facts through the serving layer; the bare weights are not a reliable source of THOX facts.
Highlights
- THOX fine-tune: a LoRA fine-tune (rank 32) on a THOX-authored corpus of 982 chat, tool-use, and factuality examples, merged into the base weights.
- GATE49 47/49 on the THOX factuality gate (development set, abstention-weighted), mean 4.837/5, 1 hard failure(s) — printed in full under Evaluation.
- This repo carries the merged weights (safetensors), a Q4_K_M GGUF, and an Ollama Modelfile.
Specifications
| Base model | Qwen/Qwen2.5-0.5B-Instruct |
| Parameters | 0.5B |
| Formats | Safetensors weights · GGUF: Q4_K_M · Ollama Modelfile |
| Training | LoRA fine-tune (rank 32) on a THOX-authored corpus of 982 chat, tool-use, and factuality examples, merged into the base weights |
| License | Apache-2.0. The THOX fine-tune and additions in this repo are covered under Apache-2.0; see the license text for terms. |
| This repository contains | Full weights and tokenizer in this repo |
Files
| File | Size | SHA-256 |
|---|---|---|
model.safetensors |
988,097,824 B | aa3b81fd4e8c92fbea25009c5488d96e5e6a743fb073c577604ffb3e2746c4d3 |
ThoxEdge-0.5B.Q4_K_M.gguf |
397,807,488 B | 6fb3222b996048fac6e4241274104382f30582b30bbddf47edcf9b56917d8598 |
tokenizer.json |
11,421,892 B | 3fd169731d2cbde95e10bf356d66d5997fd885dd8dbb6fb4684da3f23b2585d8 |
chat_template.jinja / .gitattributes / config.json / Modelfile / tokenizer_config.json / training_summary.json / generation_config.json |
— | Supporting files (see repo) |
Run it
This repo carries the merged weights (safetensors), a Q4_K_M GGUF, and an Ollama Modelfile.
ollama pull Thox-ai/thoxedge-0.5b
ollama run Thox-ai/thoxedge-0.5b
hf download Thox-ai/ThoxEdge-0.5B ThoxEdge-0.5B.Q4_K_M.gguf --local-dir ./ThoxEdge-0.5B
Platforms
| Platform / device | Run path |
|---|---|
| Ollama (laptop / desktop) | ollama run Thox-ai/thoxedge-0.5b |
| Hugging Face | Download the builds listed under Files from this repo (see Run it) |
Intended use
- On-device assistant and tool routing for ThoxMini: short conversational replies, THOX house voice, and routing intents to device and mesh functions, fully offline, with facts supplied by the serving layer.
- Out of scope: safety-critical, medical, or legal decisions without a human in the loop, and any use as a standalone source of THOX facts.
Training data and procedure
THOX fine-tune of Qwen/Qwen2.5-0.5B-Instruct. The corpus and procedure are described in this card as published at release (see Summary and Evaluation); no training beyond what this card describes is claimed for this repository.
Evaluation
Measured with the THOX GATE49 factuality gate (49 questions, weights only, no serving-side grounding; GATE49 is a development set, abstention-weighted). Full narrative under Evaluation below.
| Benchmark | Result |
|---|---|
| Total | 47/49 — mean 4.837/5 |
| Hard failures | 1 |
THOX GATE49 factuality gate (49 questions, weights only, no serving-side grounding), most recent evaluation of the shipped build: 47/49 passed, mean 4.837/5, 1 hard failure (the pricing question was answered without the required decline). THOX gate policy requires zero hard failures, so this build does not pass the gate; it is published with serving-side grounding required, which is the configuration its results support.
Limitations
- A 0.5B model: narrow capability by design. The fine-tune is a facts and behaviour skin, not a capability upgrade; expect correct declines on hard questions rather than correct answers.
- Factual answers from the bare weights remain unreliable in the weak categories; deploy with serving-side grounding for facts.
- This build has not been qualified on ThoxMini device hardware.
License
Apache-2.0. The THOX fine-tune and additions in this repo are covered under Apache-2.0; see the license text for terms.
THOX.ai — Your AI. Your Data. Your Rules.™
Provenance
Published by THOX.ai as Thox-ai/ThoxEdge-0.5B. Upstream base: Qwen/Qwen2.5-0.5B-Instruct. This card was conformed to THOX Model Card Standard v1.0 on 2026-10-31 (frontmatter, tags, header, files, and section structure); the technical content is as published at release.
Citation
@misc{thoxai2026thoxedge05b,
title={ThoxEdge-0.5B},
author={THOX.ai},
year={2026},
url={https://huggingface.co/Thox-ai/ThoxEdge-0.5B}
}