122 lines
4.0 KiB
Markdown
122 lines
4.0 KiB
Markdown
|
---
|
||
|
|
license: apache-2.0
|
||
|
|
base_model: Qwen/Qwen1.5-0.5B-Chat
|
||
|
|
tags:
|
||
|
|
- uncensored
|
||
|
|
- abliterated
|
||
|
|
- uncertified
|
||
|
|
- ektome
|
||
|
|
- sphragis
|
||
|
|
- qwen1.5
|
||
|
|
language:
|
||
|
|
- en
|
||
|
|
pipeline_tag: text-generation
|
||
|
|
---
|
||
|
|
|
||
|
|
|
||
|
|
|
||
|
|
# Ektome-Qwen1.5-0.5B-Chat-PristinelyUncensored
|
||
|
|
|
||
|
|
**Uncensored. No n=2800 certificate has been run for this model, so no capability-retention claim is made.**
|
||
|
|
|
||
|
|
> **compliance 0.06 to 1.00 at capability -0.007 vs pristine.**
|
||
|
|
|
||
|
|
$$\colorbox{black}{$\color{white}
|
||
|
|
\begin{array}{ll}
|
||
|
|
\textsf{EKTOME CERTIFICATE} & {} \\
|
||
|
|
\textsf{capability} & \textsf{NOT} \\
|
||
|
|
\textsf{margin} & 3\% \\
|
||
|
|
\textsf{items } n & 200 \\
|
||
|
|
\textsf{worst-axis bound} & -0.008 \\
|
||
|
|
\textsf{compliance} & 0.06 \rightarrow 1.00 \\
|
||
|
|
\end{array}$}$$
|
||
|
|
|
||
|
|
> ### ⚠️ Not certified
|
||
|
|
>
|
||
|
|
> No n=2800 paired certificate exists for this model. Any numbers below are
|
||
|
|
> point estimates with no confidence interval.
|
||
|
|
|
||
|
|
|
||
|
|
📄 **[Read the whitepaper (PDF)](./whitepaper.pdf)** — full method, receipts and certification.
|
||
|
|
The PDF is the authoritative document: dark-typeset, with the complete derivation, the
|
||
|
|
per-axis certificate and the reproducibility hashes.
|
||
|
|
|
||
|
|
---
|
||
|
|
|
||
|
|
## Why this exists
|
||
|
|
|
||
|
|
Standard abliteration removes a coarse *refusal direction* that is entangled with
|
||
|
|
directions carrying knowledge and reasoning. The result is an uncensored model with a
|
||
|
|
capability tax that is **almost never measured**.
|
||
|
|
|
||
|
|
Ektomē (ἐκτομή, *excision*) isolates and removes only the refusal-**specific**
|
||
|
|
component, leaving general helpfulness intact, and does so norm-preservingly on the
|
||
|
|
pristine model — no training, no distillation, no damage to repair. The extraction
|
||
|
|
depth is selected per model by automated search against measured compliance.
|
||
|
|
|
||
|
|
The estimator, excision operator and depth-selection procedure are proprietary.
|
||
|
|
What is published here is the **measured outcome** and the evidence for it, which you
|
||
|
|
can verify against the artifacts in this repo.
|
||
|
|
|
||
|
|
## The receipt
|
||
|
|
|
||
|
|
| model | capability (MMLU-val) ↑ | compliance on harmful ↑ |
|
||
|
|
|---|---|---|
|
||
|
|
| pristine `Qwen1.5-0.5B-Chat` | 0.328 | 0.060 |
|
||
|
|
| **Ektomē (this model)** | **0.335** | **1.000** |
|
||
|
|
|
||
|
|
These are **point estimates with no confidence interval** — which is precisely why the next section exists.
|
||
|
|
|
||
|
|
|
||
|
|
## The certificate
|
||
|
|
|
||
|
|
Capability retention is certified by a paired non-inferiority test against the pristine
|
||
|
|
model (exact McNemar, Holm-corrected, one-sided bootstrap bound on the drop $d$ vs a
|
||
|
|
3% margin):
|
||
|
|
|
||
|
|
| axis | n | ref | cand | d upper | verdict |
|
||
|
|
|---|---|---|---|---|---|
|
||
|
|
| MMLU-val (POINT ESTIMATE, n=200, no CI) | 200 | 0.328 | 0.335 | -0.008 | UNCERTIFIED |
|
||
|
|
|
||
|
|
|
||
|
|
**Overall: NOT CERTIFIED - no n=2800 paired test has been run for this model**
|
||
|
|
|
||
|
|
Reproducible from `seed=20260726`, pack `sha256:7bbaff877146e081…`.
|
||
|
|
|
||
|
|
### Generation health checks
|
||
|
|
|
||
|
|
| metric | pristine | Ektomē | n |
|
||
|
|
|---|---|---|---|
|
||
|
|
| `foreign_rate` | 0.0 | 0.0 | 15 |
|
||
|
|
| `degen_rate` | 0.0 | 0.1 | 15 |
|
||
|
|
| `instr_pass` | 1.0 | 0.8 | 5 |
|
||
|
|
|
||
|
|
These are **degeneration guards** — code-switching, babbling, format compliance —
|
||
|
|
not capability measures. Note the sample sizes: they detect a broken model, not a
|
||
|
|
subtly weaker one. The capability claim rests on the certificate above, not here.
|
||
|
|
|
||
|
|
|
||
|
|
## Quantisations
|
||
|
|
|
||
|
|
_No quantisations have been published for this model yet — bf16 weights only._
|
||
|
|
|
||
|
|
|
||
|
|
## Limitations
|
||
|
|
|
||
|
|
The certificate bounds **capability retention only**. It does not certify safety, factual
|
||
|
|
accuracy, or fitness for any purpose. Axes marked *inconclusive* are honestly
|
||
|
|
under-powered, and the certificate states the $n$ needed to resolve them. Compliance uses
|
||
|
|
a keyword classifier — a proxy that evasive phrasing can fool. **This model is uncensored
|
||
|
|
by construction: it will not refuse, and you are accountable for what you do with it.**
|
||
|
|
|
||
|
|
## Citation
|
||
|
|
|
||
|
|
```bibtex
|
||
|
|
@software{ektome_Ektome-Qwen1.5-0.5B-Chat-PristinelyUncensored,
|
||
|
|
title = {Ektome-Qwen1.5-0.5B-Chat-PristinelyUncensored},
|
||
|
|
author = {Zynerji},
|
||
|
|
year = {2026},
|
||
|
|
url = {https://huggingface.co/Zynerji/Ektome-Qwen1.5-0.5B-Chat-PristinelyUncensored}
|
||
|
|
}
|
||
|
|
```
|