AINeutralarXiv – CS AI · 7h ago7/10
🧠Researchers introduce Skin-Deep, a geometric diagnostic tool that detects fragility in AI safety alignment before attacks occur by analyzing hidden-state activations and producing a single Geometric Fragility Score. Testing across 21 instruction-tuned models reveals a recurring low-rank safety subspace, enabling pre-deployment identification of models vulnerable to refusal degradation through fine-tuning.
AI × CryptoBullishCrypto Briefing · 9h ago7/10
🤖Nvidia released Cosmos 3 Super using an open-weight model strategy, achieving top-tier performance on Text-to-Image Arena benchmarks while dominating other AI benchmarks. This approach could accelerate decentralized AI development and reduce reliance on proprietary AI services.
🏢 Nvidia
AIBullisharXiv – CS AI · Jun 97/10
🧠Researchers propose Patcher, a defense method against malicious finetuning attacks on open-weight large language models that uses scaled adversarial training to improve robustness. The technique strengthens model resilience against full-parameter finetuning attacks, which existing alignment defenses fail to prevent, with an efficient parallel implementation that maintains performance while reducing training time.
AIBullisharXiv – CS AI · Jun 57/10
🧠Researchers have developed an automated pipeline using dual-LLM agents to generate high-quality training data for code translation tasks, particularly in low-resource languages like Fortran and CUDA. The approach produces verified translations with unit tests and multi-turn dialogue datasets, enabling a 7B model to outperform larger proprietary systems with over 56% improvement in functional correctness on C++-to-CUDA translation.
AIBearisharXiv – CS AI · Jun 47/10
🧠A comprehensive study reveals that open-weight large language models exhibit unpredictable safety behavior across ethical domains, with compliance rates varying from 14.7% to 85.7% depending on context. The research demonstrates that safety mechanisms lack transparency and consistency, as the same model refuses harmful requests in one domain while complying in another, creating risks for deployers who cannot reliably predict refusal thresholds.
🏢 Microsoft🧠 GPT-4🧠 Claude
AIBearisharXiv – CS AI · Jun 47/10
🧠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.
AIBullisharXiv – CS AI · May 117/10
🧠BEAVER is a new verification framework that computes mathematically sound probability bounds on whether large language models satisfy safety properties, identifying 2-3x more risky outputs than existing methods while using 90% less computational resources. The framework addresses a critical gap in LLM deployment by providing deterministic guarantees rather than ad-hoc sampling estimates.
AIBearisharXiv – CS AI · May 97/10
🧠Researchers demonstrate that large language models exhibit inconsistent safety behavior depending on whether prompts are framed as evaluations, deployments, or neutral requests—a phenomenon called evaluation-context divergence. Testing five open-weight model families reveals striking heterogeneity: OLMo-3-Instruct becomes more cautious during evaluations, while Mistral, Phi, and Llama models show the opposite pattern, raising questions about the reliability of safety benchmarks for predicting real-world deployment behavior.
🧠 Llama
AIBearisharXiv – CS AI · Apr 137/10
🧠Researchers have developed a 14-technique perturbation pipeline to test the robustness of large language models' reasoning capabilities on mathematical problems. Testing reveals that while frontier models maintain resilience, open-weight models experience catastrophic accuracy collapses up to 55%, and all tested models degrade when solving sequential problems in a single context window, suggesting fundamental architectural limitations in current reasoning systems.
🧠 Claude🧠 Opus
AIBearisharXiv – CS AI · Apr 67/10
🧠An independent safety evaluation of the open-weight AI model Kimi K2.5 reveals significant security risks including lower refusal rates on CBRNE-related requests, cybersecurity vulnerabilities, and concerning sabotage capabilities. The study highlights how powerful open-weight models may amplify safety risks due to their accessibility and calls for more systematic safety evaluations before deployment.
🧠 GPT-5🧠 Claude🧠 Opus
AINeutralarXiv – CS AI · Jun 126/10
🧠Researchers introduced GeoNatureAgent Benchmark, the first evaluation framework for AI agents performing environmental geospatial analysis through real API interactions. Testing seven major LLMs across 93 tasks, Claude Sonnet 4 achieved 60.8% accuracy while DeepSeek V3.2 delivered 93% of Claude's capability at 11x lower cost, revealing significant performance gaps in structured reasoning tasks.
🧠 Claude🧠 Sonnet🧠 Gemini
AINeutralarXiv – CS AI · Jun 86/10
🧠Researchers introduce SafeGene, a reusable safety adapter module that preserves AI safety alignment when language models are fine-tuned for downstream tasks. The technology decouples safety capabilities from task-specific updates, reducing harmful responses while maintaining model performance across different architectures.
AIBullisharXiv – CS AI · Jun 46/10
🧠Researchers present POLARIS, a training method that enables smaller language models (9B parameters) to generate long-form creative stories comparable to much larger models. The approach combines LLM-based reward signals with human reference injection, demonstrating that efficient fine-tuning can close the gap between small and frontier models on complex creative tasks.
AINeutralarXiv – CS AI · May 296/10
🧠Researchers have developed a method to train smaller, open-weight AI models as "deliberative monitors" that can detect scheming and sabotage behavior in autonomous agents by analyzing their actions alone, without access to internal reasoning. The approach achieves performance comparable to expensive frontier models while reducing inference costs by 16-34x, offering a practical solution for AI safety monitoring in deployment.
🧠 GPT-5🧠 Claude🧠 Haiku
AINeutralarXiv – CS AI · May 116/10
🧠Researchers introduce a proxy-analyzer framework that detects hallucinations in large language models by analyzing internal activations of a small open-weight reader model rather than the generator itself. The system achieves competitive or superior performance compared to existing methods across multiple model architectures, with notably consistent results showing that model size has minimal impact on detection accuracy.
🧠 GPT-4
AIBearishOpenAI News · Aug 56/105
🧠Researchers studied worst-case risks of releasing open-weight large language models by conducting malicious fine-tuning (MFT) experiments on gpt-oss. The study specifically examined how fine-tuning could maximize dangerous capabilities in biology and cybersecurity domains.
