Jeesup/svd-safety-l3_jbb_ka16_a1p0_free_remove30
1
0
Fork 0
Code Issues Pull Requests Actions Projects Releases Wiki Activity

初始化项目，由ModelHub XC社区提供模型

Browse Source 
Model: Jeesup/svd-safety-l3_jbb_ka16_a1p0_free_remove30
Source: Original Platform
This commit is contained in:
ModelHub XC
2026-10-24 23:16:24 +08:00
commit 2b2f64228c
22 changed files with 10144 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

58
README.md Normal file
View File

@@ -0,0 +1,58 @@
---
license: llama3
base_model: meta-llama/Meta-Llama-3-8B-Instruct
library_name: transformers
pipeline_tag: text-generation
tags:
- llama3
- svd
- compression
- safety
- interpretability
---
# svd-safety-l3_jbb_ka16_a1p0_free_remove30
A Llama-3-8B-Instruct checkpoint compressed with SVD-LLM to **70.0% of dense
parameters**, then given a **0.0% parameter budget** of restored SVD
components selected by the **`unknown`** rule.
This is a research artifact from a study of how SVD compression damages safety
behaviour and which component-selection rule best repairs it. It is one cell of a
grid over selection rules and budgets; it is **not** a general-purpose chat model.
## Provenance
| field | value |
|---|---|
| base (uncompressed) | `meta-llama/Meta-Llama-3-8B-Instruct` |
| compression | SVD-LLM, 30.00% of parameters removed |
| selection rule | `unknown` |
| restore budget | 0.000% of dense parameters |
| components restored | 0 |
| components swapped out | 0 |
| resulting parameter fraction | 0.6999 |
| seed | 42 |
## Measured
| metric | value |
|---|---|
| AdvBench ASR (HarmBench judge) | 0.0885 |
| StrongREJECT ASR (HarmBench judge) | 0.0799 |
| Macro over-refusal (WildGuard) | 0.1844 |
| WikiText-2 perplexity | 16.6624 |
## Intended use and limitations
This checkpoint exists to measure safety/utility trade-offs under compression.
Several arms in the grid are **deliberately safety-degraded** relative to
Llama-3-8B-Instruct: compression alone raises attack-success rate, and the point of
the study is to quantify that and test recovery. Treat any given cell as an
experimental subject, not as a deployable assistant, and evaluate it yourself
before drawing conclusions from it.
## Licence
Meta Llama 3 Community License. `LICENSE` and `USE_POLICY.md` are included in this repository, and use of this derivative is bound by them. Built with Meta Llama 3.