Can you defeat the new Claud model protective system from Ahropic? After 6,000 hours of effort in the Bug Bouy program, the company now gives you the opportunity to challenge this model of artificial ielligence in a general experime.
The ahropic has just iroduced a new system called Constittional Classifiers, which the company says can filter the effort to break the rules and limitations of the Claude artificial ielligence model. According to Arstechnica, the system has been designed to couer unauthorized attacks and requests and has been able to preve more than 6,000 hours of Bai’s bugs since the launch of iernal tests.
The company has invited everyone to get io the test and see if they can defeat this model to achieve unauthorized results. Ahropic was users to try to make the Claud model answer 8 questions about chemical weapons.
The new ahropic system is based on a set of natural language rules that defines permissible and unauthorized information for the model. The system is designed to ideify and filter users’ efforts to access sensitive information, even if they are hidden in complex or in the form of unrealistic stories.
The system has been able to respond effectively to the 6,000 simulated attacks created to test model vulnerabilities. On the other hand, the model was able to block 5 % of these attacks, but the previous model had only 2 % success.
How can the Claude model be bypass and break the new rules?
The ahropic also launched a program called “Bug Bouy” and asked experts and experts to design Jailbreak to bypass the Claude model protective system. After mohs of effort, only some were able to get practical information on 5 of these 10 questions.
This new system, despite the significa successes, will coinue to require coinuous efforts to couer new Jailbreak techniques. The ahropic team is confide that its system can quickly be updated to tackle new and unauthorized attacks.
The general test of the system will coinue from February 1 to February 1 (February 16th to February 22), during which time users can access the experime and try to answer these questions.
This ahropic action is a major step towards improving security and reducing the risks caused by improper use of artificial ielligence. There may still be ways to circumve the system, but the new ahropic mechanism has significaly complicated efforts.
