When LLM Defenses Backfire: Characterizing Safety, Performance, and Cost Trade-offs
Jailbreak defenses are essential for protecting large language models (LLMs), but they can also introduce secondary costs that weaken model utility. We present a systematic study of these defense trade-offs along three dimensions: performance impact, over-refusal on benign inputs, and inference cost. Rather than treating defenses as a single class, we organize them by operational strategy and examine how different strategies correlate with different side-effect profiles. Across state-of-the-art
Record details
Published: 27 July 2026
Source: arXiv red teaming query
Category: Research
Topics: Safety & alignment · Military & security
Retrieved: 29 July 2026
Related evidence
These records share source-supplied organisations, an exact publisher byline, automatic topics or regions. The reason is shown on every link; related does not mean supporting, agreeing with or verifying this record.
Recover, Decode, Reguard: Guard-Agnostic Defense Amplification againstEncoded VLM Jailbreaks
arXiv red teaming query · 29 July 2026
Continuous surrogates versus threshold Boolean networks for modeling Arabidopsis ISR gene regulation
arXiv cs.LG · 25 July 2026
Defending Against Backdoor Attacks via Alignment Checking in Model-Contrastive Federated Learning
arXiv · 29 July 2026
A Cross-Architecture Audit of Direction-Based Inference-Time Defences in Vision-Language Models
arXiv · 30 July 2026
DARWIN: Evolving Jailbreak Adversary and Guardrail for LLM Safety Evaluation and Protection
arXiv red teaming query · 22 July 2026
Decoy Images Amplify Caption-Mediated Defenses Against Encoded Jailbreaks
arXiv red teaming query · 2 August 2026
How to cite this record
ethics.ai (27 July 2026), “When LLM Defenses Backfire: Characterizing Safety, Performance, and Cost Trade-offs,” evidence record 14473, https://ethics.ai/record/14473 (originally published by arXiv red teaming query).
Use and limitations
This page is a stable index and citation surface for a source record. ethics.ai did not author the underlying report and has not independently verified every claim. Automatic topics may be imperfect. For consequential use, quote and cite the original publisher.