Welcome to a Wednesday edition where AI is finally solving the mysteries that have long stumped even the best human engineers.
MODELS
🤖 Claude Fable 5.1 and Mythos 5.1 launch with tiered safeguards
Image source: anthropic.com
Claude Fable 5.1 found the cause of a rare crash in Millennium's internal systems that none of its engineers had been able to explain after several years of trying.
The details:
Claude Fable 5.1 and Claude Mythos 5.1 are the same model, but with different levels of safeguards.
Fable 5.1 is generally available, while Mythos 5.1 is available only through trusted access programs.
Fable 5.1 will cost an estimated 25% less than Fable 5 for typical workloads, wherever usage is billed by token.
OpenAI shared new details on its forthcoming Astra model. OpenAI said Astra is the first large language model to meet its critical cybersecurity threshold.
The details:
OpenAI determined that Astra is capable of finding unknown security flaws in computer systems and exploiting them without a person's guidance.
OpenAI noted that Astra scored a perfect score on ExploitBench.
In a modified version of the test developed by OpenAI engineers, Astra discovered and exploited two zero-day vulnerabilities.
Large Vision-Language Models (LVLMs) remain prone to hallucinations. Part of this failure is connected to a measurable property of their representations called feature instability.
The details:
Hallucinations produce responses that are irrelevant or inconsistent with the multimodal input.
Existing mitigation methods mainly rely on external supervision, output calibration, or attention regulation.
The internal representation dynamics of autoregressive generation remain underexplored.
