memo-ozdincer/jepa-qwen3-32b-ce-cb-pra-2026-05-20
Qwen3-32B — CE-floor / Circuit-Breaker × Identity / MLP PRA Trained LoRA adapters + attack results + benign evals for the Qwen3-32B PRA matrix (samuelsimko/AdversariaLLM experiments/configs/qwen3_32b_ce_cb.yaml @ commit 3a21c23), executed on a single 8×H200 tamia node. Paper context: Predictive Representation Alignment Improves Generalization in LLM Safety (work in progress). PRA = a small predictor head that maps adversarial-view hidden states toward the clean view's… See the full description on the dataset page: https://huggingface.co/datasets/memo-ozdincer/jepa-qwen3-32b-ce-cb-pra-2026-05-2
| Author / organization | memo-ozdincer |
|---|---|
| License (as declared) | other |
| Downloads | 249 (as reported by Hugging Face) |
| Likes | 0 |
| Created | 21 May 2026 |
| Last updated at source | 21 May 2026 |
Popularity, last 3 days
249 now (+3 since 2026-10-16). Counts as reported by the source, snapshotted daily by Silytics.
Recent changes
- 17 Oct 2026 Description: Qwen3-32B — CE-floor / Circuit-Breaker × Identity / MLP PRA Trained LoRA adapters + attack results + benign evals for the Qwen3-32B PRA matrix (samuelsimko/AdversariaLLM experiments/configs/qwen3_32b_ce_cb.yaml @ commit 3a21c23), executed o → Qwen3-32B — CE-floor / Circuit-Breaker × Identity / MLP PRA Trained LoRA adapters + attack results + benign evals for the Qwen3-32B PRA matrix (samuelsimko/AdversariaLLM experiments/configs/qwen3_32b_ce_cb.yaml @ commit 3a21c23), executed o
Changes detected by Silytics between refreshes of the source.
More from memo-ozdincer
Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 18 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.