# Attribution and provenance

## Upstream model

- Project: `Qwen/Qwen3.8-27B`
- Immutable revision: `1d4bf0f2ff6012fd82039f2fa52739d0dd7c60c0`
- License: Apache-2.0
- Source: <https://huggingface.co/Qwen/Qwen3.8-27B/tree/1d4bf0f2ff6012fd82039f2fa52739d0dd7c60c0>

The BF16 GGUF input was the separately pinned conversion
`ggml-org/Qwen3.8-27B-GGUF@0669b98607d47046c7c2b3f801011d54a08cfccf`.

## External calibration and importance matrix

The rendered calibration corpus and importance matrix were produced and
published by Bartowski (bartowski1182), not by this project:

- Repository: `bartowski/Qwen3.8-27B-GGUF`
- Immutable revision: `f0eec4a4bb4975114a030d048952d83c0a53c034`
- Importance matrix: `Qwen3.8-27B-imatrix.gguf`
- SHA256: `aaa933d4b9ce23e1f65c548ad34f16956d8af44a51b5c15bf4f393ba59508cd8`
- Rendered corpus SHA256: `b3e9d2371c8931cafa5b6184adcee5cb59adb7f4103e28d0f99eddac1cced4e6`
- Source-data gist revision: `14d426cb41b6898129b2904bdafeca8ba56ff88e`
- Tool-reasoning source dataset: `interstellarninja/hermes_reasoning_tool_use` (Apache-2.0)

This attribution does not imply that Bartowski validated or endorsed this
mixed-quant artifact.

## Quantization toolchain

- Project: `ggml-org/llama.cpp`
- Release: `b10442`
- Commit: `9b0a2ce859d3884252705e9ac7c93c98616bb238`

The toolchain is not included in the upload set. Its own license remains with
its source distribution.