# chujiezheng/Mistral7B-PairRM-SPPO-ExPO

This model is an extrapolated alignment model created from SFT and DPO or RLHF checkpoints, as described in the paper reference on weak to strong extrapolation. It is described as aligning with human preference.

## Cite this

- Page: https://huggingbay.xyz/artifact/hf-model-chujiezheng-mistral7b-pairrm-sppo-expo
- Citation: chujiezheng/Mistral7B-PairRM-SPPO-ExPO. The Hugging Bay model record. https://huggingbay.xyz/artifact/hf-model-chujiezheng-mistral7b-pairrm-sppo-expo (retrieved 2026-11-22).
- Check a downloaded file: run `shasum -a 256 <file>` (macOS), `sha256sum <file>` (Linux) or `Get-FileHash -Algorithm SHA256 <file>` (Windows), then compare with that file's SHA-256 at https://huggingbay.xyz/api/artifacts/hf-model-chujiezheng-mistral7b-pairrm-sppo-expo/files?limit=100. How checks work: https://huggingbay.xyz/verification

## Facts

- Type: LLM
- Task: text-generation
- Framework: transformers
- License: apache-2.0: Permissive open license: commercial use, modification, and redistribution are permitted. (https://spdx.org/licenses/Apache-2.0.html)
- Source: https://huggingface.co/chujiezheng/Mistral7B-PairRM-SPPO-ExPO
- Hosting: hosted (9 files)
- Files: 9 hosted of 12 listed
- File counts: "listed" is every file in the source repository; only hosted files download from Hugging Bay, and the download plan lists the files it selects for one download, so its count can differ.
- Hosted here: 9 hosted of 12 listed
- Full repo upstream: 14.5 GB
- Recommended file: 14.5 GB (3 files)
- Parameters: 7B
- Download: https://huggingbay.xyz/api/artifacts/hf-model-chujiezheng-mistral7b-pairrm-sppo-expo/download-plan (download plan: https://huggingbay.xyz/api/artifacts/hf-model-chujiezheng-mistral7b-pairrm-sppo-expo/download-plan)
- Per-file SHA-256: available for 9 of 9 hosted files via https://huggingbay.xyz/api/artifacts/hf-model-chujiezheng-mistral7b-pairrm-sppo-expo/files?limit=100
- Availability: Ready
- Downloadable: yes
- Run-ready: no (Hugging Bay has not certified a run path for this model; hosted bytes are downloadable and hash-checkable only)
- Peer-assisted fallbacks: 0 verified, 0 pending review
- Downloads: 7889
- Stars: 0
- Card completeness: 96%

## Description

This model is an extrapolated alignment model created from SFT and DPO or RLHF checkpoints, as described in the paper reference on weak to strong extrapolation. It is described as aligning with human preference.

## Model Card Facts

- Declared license: apache-2.0
- Pipeline: text-generation
- Library: transformers
- Task categories: text-generation
- Languages: en
- Paper ids: 2404.16792
- Model-index entries: Mistral7B-PairRM-SPPO-ExPO
- Evaluation signals: 6 model-card eval results
- File hints: 12 upstream file samples

## Evaluation Signals

- IFEval (0-Shot): strict accuracy = 36.73 (Text Generation)
- BBH (3-Shot): normalized accuracy = 13.68 (Text Generation)
- MATH Lvl 5 (4-Shot): exact match = 0.91 (Text Generation)
- GPQA (0-shot): acc_norm = 3.58 (Text Generation)
- MuSR (0-shot): acc_norm = 8.66 (Text Generation)
- MMLU-PRO (5-shot): accuracy = 17.24 (Text Generation)

## Referenced Data

- Paper: 2404.16792

## File Hints

- .gitattributes (1.5 KB)
- README.md (6.8 KB)
- config.json (650 B)
- generation_config.json (132 B)
- model-00001-of-00003.safetensors (4.6 GB) LFS
- model-00002-of-00003.safetensors (4.7 GB) LFS
- model-00003-of-00003.safetensors (4.2 GB) LFS
- model.safetensors.index.json (23 KB)
- special_tokens_map.json (551 B)
- tokenizer.json (1.7 MB)
- tokenizer.model (482 KB) LFS
- tokenizer_config.json (1.4 KB)

## Quality Checklist

- Summary: passed
- Description: passed
- Model card metadata: passed
- License: passed
- Provenance: passed
- File inventory: passed
- Hosted bytes: passed
- Scan result: passed
- Manifest signature: passed

## Safety Notes

- External catalog metadata imported from Hugging Face public API on 2026-09-21T22:10:24.193Z.
- Open license detected as apache-2.0.
- Model-index metadata found for Mistral7B-PairRM-SPPO-ExPO.
- Not a run kit: hosted, hash-checked bytes only; run readiness is not certified for this model yet.

## Actions

- Web: https://huggingbay.xyz/artifact/hf-model-chujiezheng-mistral7b-pairrm-sppo-expo
- Files API: https://huggingbay.xyz/api/artifacts/hf-model-chujiezheng-mistral7b-pairrm-sppo-expo/files?limit=100
- Metadata: https://huggingbay.xyz/api/artifacts/hf-model-chujiezheng-mistral7b-pairrm-sppo-expo/metadata.json
- Distribution: https://huggingbay.xyz/api/artifacts/hf-model-chujiezheng-mistral7b-pairrm-sppo-expo/distribution
- Manifest: https://huggingbay.xyz/api/manifests/hf-model-chujiezheng-mistral7b-pairrm-sppo-expo
- Download plan (per-file SHA-256 + signed download URLs): https://huggingbay.xyz/api/artifacts/hf-model-chujiezheng-mistral7b-pairrm-sppo-expo/download-plan
- Hosted download rows: https://huggingbay.xyz/api/downloadable-files?artifactId=hf-model-chujiezheng-mistral7b-pairrm-sppo-expo&limit=100
- Request mirror: not applicable.
- Submit peer fallback for review: POST https://huggingbay.xyz/api/artifacts/hf-model-chujiezheng-mistral7b-pairrm-sppo-expo/decentralized-fallbacks
