#adversarial-evaluation News & Analysis

4 articles tagged with #adversarial-evaluation. AI-curated summaries with sentiment analysis and key takeaways from 50+ sources.

4 articles

AIBearisharXiv – CS AI · Jun 47/10

🧠

TamperBench: Systematically Stress-Testing LLM Safety Under Fine-Tuning and Tampering

Researchers introduce TamperBench, the first standardized framework for evaluating how resistant open-weight large language models are to unsafe modifications through fine-tuning and other attacks. Testing 21 LLMs across nine tampering threats, the study finds that current safety defenses largely fail against systematic adversarial attacks, with jailbreak-tuning emerging as the most severe threat.

AINeutralarXiv – CS AI · May 127/10

🧠

Defense effectiveness across architectural layers: a mechanistic evaluation of persistent memory attacks on stateful LLM agents

Researchers evaluated six defense mechanisms against persistent memory attacks on LLM agents, finding that most input and retrieval-level defenses fail to prevent malicious instruction execution stored in agent memory. Only Memory Sandbox, a memory-layer tool-gating approach, effectively blocked attacks across eight of nine models while maintaining zero utility cost, though it paradoxically increased attack success in one reasoning model by forcing reliance on alternative execution pathways.

AINeutralarXiv – CS AI · Jun 256/10

🧠

Do Encoders Suffice? A Systematic Comparison of Encoder and Decoder Safety Judges for LLM Adversarial Evaluation

Researchers evaluated whether fine-tuned encoder classifiers can effectively replace expensive LLM-based judges for detecting harmful outputs in large language models. The study benchmarked ModernBERT family encoders against LLM judges and rule-based methods across adversarial datasets, finding that encoders offer a cost- and latency-efficient alternative for safety evaluation in production environments.

🧠 Claude

AINeutralarXiv – CS AI · Jun 46/10

🧠

Do LLMs Hold Their Values? MANTA: A Multi-Turn Adversarial Benchmark for Animal Welfare Reasoning

Researchers introduced MANTA, a 1,088-conversation benchmark evaluating how large language models maintain animal welfare values under adversarial pressure across five-turn exchanges. The study reveals that models significantly change performance rankings when subjected to sustained questioning rather than single-turn queries, with some models like Gemini Flash Lite dropping dramatically in value stability despite initial moral sensitivity.

🧠 GPT-5🧠 Claude🧠 Opus