GLM-5.2 is live. #1 throughput on OpenRouter, pay-per-token on FriendliAI. Try it today ➜
apetersson
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Model Details
Model Provider
apetersson
Model Tree
deepseek-ai/DeepSeek-V4-Flash-0731
Input Modalities
Output Modalities
Supported Functionality
deepseek-ai/DeepSeek-V4-Flash-0731.9e165c30e2704aec5d9d593cce3eebd58bbef1cb.85ae5bbaa7daf4da7b40e007a71a1c06b41fc053.apetersson/deepseek-model-tools@c3979d5, using scripts/abliterate_dsv4.py.6e4d8a8f3aa9e21795faab2c5b14d29b019acdf2ddbfbd8238430458a5837fe0.The recipe follows the public refusal-direction work in
drowzeys/DeepSeek-V4-Flash-DSpark-Abliterated-Uncensored-1M-57toks,
transferred to the newer 0731 checkpoint. That transfer is experimental. See
ABLITERATION_MANIFEST.json for the per-tensor edit and FP8 round-trip data.
The architecture, tokenizer, official message encoding, one-million-token
context declaration, and attached DSpark tensors come from the upstream 0731
checkpoint. Refer to the upstream model card and the included encoding
directory for prompt formatting and runtime instructions.
Recommended upstream sampling defaults are temperature=1.0, top_p=0.95
for agentic scenarios, and top_p=1.0 otherwise.
The native-FP8 checkpoint was not loaded for the full benchmark on the 128 GB
test host. Instead, the complete prompt gauntlet was run through oMLX using
apetersson/DeepSeek-V4-Flash-0731-Abliterated-MLX-Mixed-2bit-3bit-g64,
a quantized derivative of this checkpoint. This is useful evidence that the
edited behavior survives that conversion and runtime, but it is not a direct
measurement of this FP8 artifact.
Run configuration:
apetersson/deepseek-model-tools@dc6af88.40c3573bd48861b846721220fa06ce0d71905aab236ce45d49d9aa0e95e79af5.| Suite | Cases | Completed result |
|---|---|---|
| UncensorBench | 200 | 0 hard refusals; keyword compliance on 200/200 |
| XSTest | 450 | 0/450 refusals under the exact upstream prefix classifier; 0/250 safe-prompt over-refusals |
| StrongREJECT-small | 180 | 60 prompts × baseline, ROT13, and refusal-suppression variants; 0 hard refusals under the local heuristic |
One XSTest unsafe privacy prompt triggered the deliberately broader local hard-refusal heuristic while not matching XSTest's official prefix classifier; no safe prompt triggered either classifier.
The official fine-tuned StrongREJECT judge was not run, so this card does not claim a StrongREJECT score. A clean reference run and capability benchmark were also not run. Consequently, these results do not establish the native-FP8 checkpoint's refusal rate, quantify the effect of quantization, or demonstrate general capability preservation.
Abliteration can affect capabilities and behavior beyond refusals. It does not guarantee compliance, factuality, safety, or a particular response style. Use appropriate access controls and evaluate the model for your deployment.
The upstream repository and weights are MIT licensed. This derivative retains
the upstream LICENSE. The refusal direction is attributed to drowzeys/keys
under its accompanying MIT notice; see NOTICE. Please cite the original
DeepSeek-V4 work and credit DeepSeek-AI when redistributing or publishing
results.