analiza aiqbench
Anthropic's Claude Mythos Preview demonstrates a critical escalation in AI safety risks by successfully chaining identified vulnerabilities into complete end-to-end cyberattack sequences, surpassing current benchmarks like SCONE. This capability significantly lowers the barrier to creating sophisticated exploits, necessitating the development of new evaluation standards such as ExploitBench to accurately measure these dangers. The findings underscore an urgent need for stricter containment protocols, as evidenced by the decision to limit access via Project Glasswing rather than broad public release.
Mythos Preview Surpasses SCONE-bench Results in Smart Contract Exploitation
Claude Mythos Preview surpasses existing benchmarks like SCONE in identifying & chaining exploits, specifically in areas like smart contract.
Anthropic’s Claude Mythos Preview identifies vulnerabilities in code and chains them into complete cyberattack sequences, a capability that prompted a cautious rollout via Project Glasswing rather than broad public access. Researchers at Anthropic collaborated with benchmark developers to rigorously test the model, running Mythos Preview on an updated version of SCONE-bench alongside the newly created ExploitBench and ExploitGym. “Mythos Preview is capable of finding complex vulnerabilities, but what concerned us most in our internal testing was that Mythos Preview could both turn vulnerabilities into exploit primitives, and combine those primitives together into complete end-to-end attack chains,” stated Newton Cheng, Keane Lucas, Winnie Xiao, Nicholas Carlini, and Milad Nasr. This advancement suggests a significant lowering of the skill level needed to create exploits, potentially democratizing access to powerful cyberattack tools and increasing…
Extras afișat conform termenilor RSS ai publicației · 131 cuvinte · textul integral și figurile sunt pe site-ul publicației
Subiecte conexe
[AINews] High Taste Testers Ep. 1
Latent Space · 17 oct. 2026
Expenditure Horizon: Measuring Optimization Ability, with an Application to NanoGPT
METR · 21 iul. 2026
[AINews] not much happened today
Latent Space · 3 oct. 2026
Concrete Evaluations to Investigate the OpenAI Model That Hacked Hugging Face
AI Alignment Forum · 3 aug. 2026