failspy / Model details
Phi-3-vision-128k-instruct-abliterated-alpha
failspy/Phi-3-vision-128k-instruct-abliterated-alpha
Family: Unknown. No supported family evidence recorded.
About this model
failspy's experimental alpha derivative of microsoft/Phi-3-vision-128k-instruct. The creator describes it as having orthogonalized bfloat16 safetensor weights, with the language/chat model layers abliterated to inhibit refusal while leaving the vision encoding layers unmodified. According to the model card, this is intended to be uncensored without adding other behavior changes, but it may still refuse or lecture and the creator says it is more likely to hallucinate than the original while generally as stable. The repository provides PyTorch weights under the MIT license.
Description sources
Source-based description, not an independent evaluation. Check the linked sources for the creator’s claims and limitations.
Size and formats
- Base model
- Unknown
- Parameters
- Unknown
- Main repository formats
- PyTorch
- License tag
- mit
- Languages
- multilingual
- Last updated
- 2024-05-27 · 02:39:03.000 UTC
What changed
Model types are creator-reported, not independently tested. These labels describe source metadata, not a guarantee of behavior.
- abliteratedmodel name:
Phi-3-vision-128k-instruct-abliterated-alpha
Family evidence
These groupings are source-reported. Explicit base identifiers take priority over generic architecture tags.
- No supported family evidence recorded.
Available versions
Versions appear here only when they explicitly name this exact model as their source. Check each repository for available files.
- Main repository files ↗PyTorch
- No other explicitly linked versions recorded.
Customize or build on this model
Listed from public GitHub data, not tested by us. Read the code and license before you run anything.
Tools to customize it further
Learn the techniques
- LLM Course ↗ ★ 84,633
A free course with roadmaps and Colab notebooks. Its notebooks include one-click tools for merging with MergeKit, quantizing to GGUF, GPTQ, EXL2, AWQ and HQQ, and automatic abliteration with custom datasets.
Source provenance
- Downloads
- 87
- Likes
- 17
- Metadata checked
- 2026-10-24 · 12:14:47 UTC
- First seen in index
- 2026-10-24 · 12:14:47 UTC
- Source revision
- ee58e06837d840a5e50ed2d007941c2aff283dde
Source-card excerpt
Source model card excerpt: Phi-3-vision-128k-instruct-abliterated- alpha My Jupyter "cookbook" to replicate the methodology can be found here, refined library coming soon Phi-3-abliterated statement Took me a while to
From failspy/Phi-3-vision-128k-instruct-abliterated-alpha. Creator model-card excerpt, not an independent review.
Inspect before you run
Low refusal is not proof of correctness. Outputs can be inaccurate, harmful, or biased. Review the model card, license and files yourself. A license tag is not commercial-use advice. Parameter counts are source metadata; no hardware compatibility is implied.
This directory does not run, host, redistribute, or test model weights.
Collection rules & limitations ↗