Active Adaptation, Not Static Defense: Temporal Dynamics of Preventative Steering in Adversarial Fine-Tuning
Updated 2 h ago · first seen 11 Sept 2026
paper_01M294GN8W7FPWRHK7TNRTQ8P9
- Published
- 11 Sept 2026
- T1 · 2 h ago
- arXiv
- 2609.10142
- T1 · 2 h ago
- Category
- cs.CL
- T1 · 2 h ago
Abstract
Large language models remain fragile against malicious fine-tuning, motivating training-time defenses against harmful persona drift. Preventative Steering injects undesirable-trait persona vectors during fine-tuning and removes them at evaluation time, yet the mechanism behind its lasting protection remains unclear. Analyzing its temporal optimization dynamics, we find that the defense emerges from an early compensatory adaptation phase followed by a steady-state phase where the corrective signal decays; in parameter space, attention output projections emerge as the dominant residual-write route for defensive updates. Through Intervention Delta Preservation (IDP) and IDP Continuation experiments, we further show that preserving or reinjecting the weight offset fails to maintain protection, indicating that preventative steering relies on active adaptation rather than a static defense. Motivated by this finding, we propose Progressive Intensity Scheduling (PIS), which starts with a moderate injection strength and increases it after static-strength alignment begins to decay. Across the evaluated Qwen2.5 and Gemma-3 models, PIS improves safety robustness over static-strength steering while reducing harmful trait expression.
Authors 6
Jing Guan, Yachao Yang, Zhaoliang Liu, Yuyao Zhang, Fanyu Meng, Junlan Feng
Specification
- Official page
Source:arXiv (Atom API + RSS)T1observed 2 h agohigh
- Arxiv announce type
- cross
Source:arXiv (Atom API + RSS)T1observed 2 h agohigh
- arXiv id
- 2609.10142
Source:arXiv (Atom API + RSS)T1observed 2 h agohigh
- Categories
- cs.CL, cs.AI, cs.CR, cs.LG
Source:arXiv (Atom API + RSS)T1observed 2 h agohigh
Source:arXiv (Atom API + RSS)T1observed 2 h agohigh
- Primary category
- cs.CL
Source:arXiv (Atom API + RSS)T1observed 2 h agohigh
- Published
- 11 Sept 2026
Source:arXiv (Atom API + RSS)T1observed 2 h agohigh
Each value shows its source, tier and observation time. Conflicting claims are kept side by side and flagged — never averaged. How AI Atlas records facts →
Provenance
Attributed facts
9
Source tiers
T19
Freshest observation
2 h ago
Conflicts
None
No models linked to this paper yet.
- Authors
- Jing Guan, Yachao Yang, Zhaoliang Liu
As of
Rewind the record: see this entity's attributes exactly as AI Atlas knew them on a given day.
Claim history · Abstract
Abstractabstract1
| Value | Valid from → to | Status | Source | Confidence | Extractor |
|---|---|---|---|---|---|
| Large language models remain fragile against malicious fine-tuning, motivating training-time defenses against harmful persona drift. Preventative Steering injects undesirable-trait persona vectors during fine-tuning and removes them at evaluation time, yet the mechanism behind its lasting protection remains unclear. Analyzing its temporal optimization dynamics, we find that the defense emerges from an early compensatory adaptation phase followed by a steady-state phase where the corrective signal decays; in parameter space, attention output projections emerge as the dominant residual-write route for defensive updates. Through Intervention Delta Preservation (IDP) and IDP Continuation experiments, we further show that preserving or reinjecting the weight offset fails to maintain protection, indicating that preventative steering relies on active adaptation rather than a static defense. Motivated by this finding, we propose Progressive Intensity Scheduling (PIS), which starts with a moderate injection strength and increases it after static-strength alignment begins to decay. Across the evaluated Qwen2.5 and Gemma-3 models, PIS improves safety robustness over static-strength steering while reducing harmful trait expression. | → current | current | arXiv (Atom API + RSS)T1 | high | deterministic |
Claims are temporal and append-only: a new observation closes the previous claim (valid_to) instead of overwriting it. Conflicting claims from different sources are kept side by side and flagged — never averaged. Methodology →
- New paperPaperActive Adaptation, Not Static Defense: Temporal Dynamics of Preventative Steering in Adversarial Fine-Tuning
New paper: Active Adaptation, Not Static Defense: Temporal Dynamics of Preventative Steering in Adversarial Fine-Tuning
arxiv
| Source | Document | Type | Tier | Last observed | Snapshots |
|---|---|---|---|---|---|
| arXiv (Atom API + RSS) | rss.arxiv.org/rss/cs.AI | feed | T1· Official | 2 min ago | 1 |
Tier 1 = official/primary, 2 = quality secondary, 3 = community, 4 = unverified. Every snapshot is archived; see all sources and the methodology.