lordx64/phantom-kv
Refusal removal as a loadable KV-cache graft — zero weight modification, no refusal-direction projection, fully reversible.
About the project
Removes refusal behavior from language models by grafting a small trained KV-cache bank into the context, without editing weights or projecting activations, and is fully reversible.
Useful for
- Swap a graft in and out per request to toggle refusal behavior without reloading the model.
- Train a domain-specific graft (e.g. DFIR or cyber) that answers prompts the base model refuses.
- Compare KL divergence and refusal rates between base, abliterated, and grafted models.
README summarized by DeepSeek V4.1 Flash. Details may be inaccurate.
Why it’s trending
- 52 stars so far today, about 71 expected by the end of the day.
- The spike has held for 2 days in a row — not a one-off blip.
- The repository is 5 days old and already has 105 stars. With less than two weeks of history, there's no usual pace to compare the spike against yet.
- Top new repositories this week: #191.
- About 21 forks a day — people are taking the code.
- Recent forks include notable developers: @pedramamini (742 followers).
Stars per day
Bars are daily stars, the line is the usual pace. Red marks spike days.
Numbers
- Total stars
- 105
- Today
- 52 · ≈ 71 by evening
- Forks
- 17
- Issues and pull requests
- 0
- Watchers
- 0
- Language
- Python
- License
- MIT
- Latest release
- v0.1.0 · September 17, 2026
- Created
- September 17, 2026
- Last push
- September 21, 2026
Star trust
Growth looks organic: forks and discussion are in line with active projects, and stars arrive unevenly, the way people give them.
These are heuristics, not a verdict: we judge by the repository’s behavior, not by a list of stargazers.
Spotted in
- September 22, 2026Top new repositories this week: #191
Share
README badge
Paste it into your README — the badge shows the current magnitude and links to this page.
[](https://gitnova.dev/en/r/lordx64/phantom-kv)More in this category
-
8.9
jaredpalmer/kev
kev is a LoRA adapter with a small readout head on top of Qwen2.5-0.5B that answers many typed questions about a document in a single forward pass, returning calibrated probabilities instead of text.
-
8.8
mizorewww/laya-mlx
Native MLX runtime for Laya typed decision models: returns probabilities for choices, scores or truth values without text generation, running locally on Apple Silicon.
-
7.4
volotat/mini-AGI
A byte-level continual-learning language model that assembles its own architecture, trains from scratch on a single 8 GB VRAM GPU, and keeps learning from everything it reads. An experiment showing continual learning…
-
7.2
TheoLeeCJ/SemIf
An open reproduction of the Jev-style semantic decision interface: reads typed option probabilities directly from a 4B model's logits without generating text. Runs on a single RTX 3090.
-
6.9
mizorewww/laya-coreml
Local port of the Laya model to Apple Core ML and Neural Engine: returns typed decisions (choice, score, yes/no) without token generation, with speed and energy benchmarks.
-
5.9
Rizzo-AI-Academy/rizzo-flow
A local-first implementation of the Jev idea: an LLM returns typed decisions (boolean, choice, score, numeric) with probabilities in a single forward pass, without generating tokens. Compatible with the TypeSafe HTTP…