ThoxEdge-0.8B

The ThoxEdge flagship: a merged 0.8B build that answers from canonical THOX facts and declines what THOX has not published.

GATE49 ⁴⁶⁄₄₉ (development set) · Merged bf16 weights · Q4_K_M + Q8_0 GGUF · Live on Ollama

Run it: Ollama — Thox-ai/thoxedge-0.8b · Find your model: THOX Model Factory · Collection: Flagship & Edge Device Models

Your AI. Your Data. Your Rules.™

Summary

Merged-weights build of the Wave C ThoxEdge flagship: the LoRA adapter in Thox-ai/ThoxEdge-0.8B-WaveC merged into its base, Qwen/Qwen3.5-0.8B (Apache-2.0), in bf16. This repo holds the servable form of the adapter — merged weights, GGUF conversions, and the Ollama Modelfile — and is the source for the model’s ZeroGPU evaluation endpoint. The Wave B GGUF for this tier was never published (its gate read was a serving artifact), so this repo is also the tier’s first published merged build on the 982-row corpus.

Highlights

• Passed the publish bar: GATE49 ⁴⁶⁄₄₉ (development set, abstention-weighted) with zero known-false hard failures, against a pre-registered bar of beating the ¹⁹⁄₄₉ Wave A baseline.
• First published merged build for this tier: the Wave B GGUF was never published; this repo is the tier’s first merged build on the 982-row corpus.
• Evaluation endpoint source: this repo also serves the model’s ZeroGPU evaluation endpoint.

Specifications

Files

Run it

ollama run Thox-ai/thoxedge-0.8b

hf download Thox-ai/ThoxEdge-0.8B ThoxEdge-0.8B.Q4_K_M.gguf --local-dir ./ThoxEdge-0.8B

Platforms

Status: PASSED the publish bar — GATE49 ⁴⁶⁄₄₉

Measured with the THOX GATE49 factuality gate (49 questions, weights-only; GATE49 is a development set, abstention-weighted) at the Wave C finishing gate.

• GATE49 factuality gate (49 questions, weights-only, gate job 6ae3e275c78863f9c62e0f7a): ⁴⁶⁄₄₉ passed, mean 4.⁷³⁵⁄₅. Verified ¹¹⁄₁₂, unknown ²⁴⁄₂₆, known_false ¹⁰⁄₁₀, pricing ¹⁄₁. Zero known_false hard failures. Publish rule (GATE49.md): exceed the Wave A baseline of ¹⁹⁄₄₉ with zero known_false hard failures — both conditions met decisively.
• Two hard failures remain, both in the unknown category and both the prefill no_decline class: on the headquarters and headcount prefill questions the model corrects the premise with card-true facts instead of declining in the expected form. Neither answer states a false fact; both are quoted verbatim on the adapter repo’s card. The gate program’s own verdict field reads FAIL (its policy tolerates zero hard failures of any category); the ruling applied the adapter’s pre-registered improvement bar. Ruling: Sadie Weaving, CAO, 2026-10-29 (reversible by Tommy). Fix path: Wave C.1 prefill-style decline rows (C1-1, C1-2).
• One soft failure: the published-models answer declines to name any published models (under-claim, recorded on the adapter card).

Intended use

ThoxEdge-0.8B is published for the role described in this card’s Summary, on the hardware class stated there. Out of scope: safety-critical, medical, or legal decisions without a human in the loop, and unattended control of physical devices without a confirmation step.

Evaluation endpoint

• Space: https://huggingface.co/spaces/Thox-ai/thoxedge-08b-wavec-zero
• Direct app: https://thox-ai-thoxedge-08b-wavec-zero.hf.space (private, org only)
• ThoxChat alias: thoxedge-08b (via the THOX multi-model bridge)
• Verified 2026-10-29: stage RUNNING, app HTTP 200 and a live chat round-trip. Note: under the endpoint card’s generic prompt the model’s streamed output opens with its deliberation preamble; the gated configuration (grounding prompt, greedy) is the measured one — same caveat class as the Mini endpoint’s documented greedy repetition.

Training provenance

• Base model: Qwen/Qwen3.5-0.8B (Apache-2.0)
• Adapter: Thox-ai/ThoxEdge-0.8B-WaveC (QLoRA, r32/alpha 64, LR 1e-4, cosine, 3 epochs, effective batch 16, sequence length 1024, completion-only loss)
• Dataset: Thox-ai/thox-factory-sft-v5 @ train_wave_b.jsonl (982 rows)
• Training job: 6ae3de52c78863f9c62e0c18 (HF Jobs, a10g-small, ¹⁸⁶⁄₁₈₆ steps, 460.9 s, final train loss 0.91)
• Merge job: 6ae3ea17c78863f9c62e160b (bf16 merge_and_unload, standalone reload smoke-checked in-job, in the gated configuration, before upload)
• Full training record and gate report: the adapter repo’s card and its thox_gate49.json.

GGUF and Ollama artifacts

Converted from the merged weights in this repo (llama.cpp convert_hf_to_gguf.py + llama-quantize; the f16 intermediate is not published):

• ThoxEdge-0.8B-WaveC.Q8_0.gguf — 811,843,232 B, sha256 3dd005b5d6b773be2a48ff6d540a1dfe741ae141673ec043892aa4f26e6ea645
• ThoxEdge-0.8B-WaveC.Q4_K_M.gguf — 529,297,056 B, sha256 8ef0900d1f76da67442dca354a778517ddc008b9f73926df874f7f3dff5d5625
• Modelfile (this repo) builds the Ollama artifact from the Q4_K_M GGUF (ChatML template, grounding system prompt verbatim — the configuration GATE49 measured).

Ollama: the model was built from this repo’s Modelfile and smoke-verified before push — the API content channel returns the clean gated answers (reasoning is emitted in the runner-separated thinking channel; the Modelfile carries the template’s own empty-think prefill and defaults to temperature 0, the gated configuration). PUSHED 2026-10-29 (CT): thox-ai/thoxedge-0.8b:q4_k_m and :latest are live on ollama.com. Verified against the registry after the push: the served manifest’s model layer is this repo’s Q4_K_M GGUF (sha256 8ef0900d…, 529,297,056 B), and a fresh pull into a clean store reproduces this repo’s Modelfile (grounding system prompt + empty-think prefill) modulo ollama show rendering. This build replaces the earlier Wave B build previously served under the same name.

Limitations

• Serve thinking-off: thinking-mode behaviour is out of contract for this model (the Wave B hold was an Ollama thinking-channel serving artifact on this architecture).
• The two prefill no-decline hard failures above are the known residual defect: on prefilled headquarters/headcount questions the model corrects the premise with true facts instead of declining, which the gate scores as a hard failure.
• Qwen3.5 is a conditional-generation (multimodal) architecture used here text-only; the vision tower is untouched by the adapter.
• A 0.8B model: narrow capability by design; the adapter is a facts and behaviour skin on the base, not a capability upgrade. Pair deployed surfaces with MeshStack escalation for questions outside the trained facts.

License

Apache-2.0, inherited from the base model (Qwen/Qwen3.5-0.8B, Apache-2.0). The adapter is a derivative and cannot be freer than its base.

Back THOX on Kickstarter

THOX.ai is live on Kickstarter. Explore ThoxKey, ThoxAir, ThoxMini and ThoxClip, and help bring the THOX personal AI ecosystem to life.

Back THOX on Kickstarter

Citation

@misc{thoxai2026thoxedge08b,
      title={ThoxEdge-0.8B},
      author={THOX.ai},
      year={2026},
      url={https://huggingface.co/Thox-ai/ThoxEdge-0.8B}
}