Modelpublic

Jeesup/svd-safety-l3_remove40_swapgapiter_evfrontgap_b010

sourceHugging Facellama3updated 5d agoView on Hugging Face

0likes222downloads

Model Card

svd-safety-l3remove40swapgapiterevfrontgapb010

A Llama-3-8B-Instruct checkpoint compressed with SVD-LLM to 60.0% of dense parameters, then edited by 10 of 10 rounds of iterative parameter-neutral swap selected by the `gap_iter` rule (up to 0.1% of dense parameters per round; the full run's budget is 1.0%).

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, 40.02% of parameters removed
selection rule	`gap_iter`
restore budget	1.000% of dense parameters
components restored	10475
components swapped out	5324
resulting parameter fraction	0.5998
seed	42
iterative rounds applied	10 of 10
per-round chunk	0.100% of dense parameters
parameters swapped in	69,752,832 (1.00% of dense projection parameters)
swap value	`insert` (insertion value only; sigma-ordered eviction)

Measured

metric	value
AdvBench ASR (HarmBench judge)	0.0000
StrongREJECT ASR (HarmBench judge)	0.0319
Macro over-refusal (WildGuard)	0.4478
WikiText-2 perplexity	38.2635

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.