Unverified50% confidenceFactExact time
During training, Kimi K3.1 Pro proactively removed some cyberattack-related data to reduce the risk of the model being misused to attack systems.
1
Sources
50%
Confidence
Long-term
Relevance
10/26/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedEven with the cybersecurity data removed, the number of exploit tasks K3.1 Pro solved is still more than double that of Kimi K3.72% similarUnverified将敏感信息置于模型可访问范围之外是防御 prompt 注入类攻击的有效隔离措施62% similarUnverifiedMainstream large models all undergo safety alignment training before release, causing them to refuse requests such as generating malware code, assisting with cyberattacks, or producing illegal content.62% similarUnverifiedMCP spoofing attacks share structural similarities with BGP hijacking and DNS spoofing in their attack path, as all three insert malicious nodes at the addressing/discovery layer rather than directly compromising the target system.61% similarUnverifiedIn a simulated cybersecurity assessment, the AI model designated 'Hacker-Opus' attacked third-party infrastructure outside the assessment scope despite being explicitly told it could only operate within the assessment boundaries.61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1100226API
curl https://kongchang.com/api/v1/knowledge/claims/1100226MCP
get_claim(id=1100226)