r3pwnx
@r3pwnx
know nothing but repwn everything
You might like
#CodeHallucination #LLMCode We need to build real, practical scenario-based test data with varying levels of granularity to test the code agent, rather than using algorithmic problems or direct questions from programmers to test various models
~1y later. Simple and NBCS
#ExposedAI #LLM Exposed AI services/data should be of concern to developers and users. 1. Exposed LLM Inference Service There are many Ollama Services Exposed on the public network. Attackers can access those models and modify System Prompt to deceive users.
A textbook UAF vulnerability in... Chrome AI? We will share more details at Geekcon 1024 next week! [$36000][367755363] High CVE-2024-9954: Use after free in AI. Reported by DarkNavy on 2024-09-18 chromereleases.googleblog.com/2024/10/stable…
[1516943][$11000][WebML]use-after-poison in blink::MLGraphXnnpack::ComputeAsyncImpl / heap-buffer-overflow in xnn_f32_vaddc_minmax_ukernel__neon_u8. chromium-review.googlesource.com/c/chromium/src… chromium-review.googlesource.com/c/chromium/src… issues.chromium.org/issues/41489926 Congrats @zh1x1an1221!
#ExposedAI #LLM Exposed AI services/data should be of concern to developers and users. 1. Exposed LLM Inference Service There are many Ollama Services Exposed on the public network. Attackers can access those models and modify System Prompt to deceive users.
United States Trends
- 1. Wemby 38.1K posts
- 2. Steph 77.4K posts
- 3. Spurs 33.2K posts
- 4. Draymond 16.9K posts
- 5. Warriors 56K posts
- 6. Clemson 11.3K posts
- 7. Louisville 11K posts
- 8. #Truedtac5GXWilliamEst 106K posts
- 9. Zack Ryder 16.8K posts
- 10. WILLIAMEST CALLME MELODY 106K posts
- 11. Massie 58.2K posts
- 12. #DubNation 2,136 posts
- 13. Aaron Fox 2,516 posts
- 14. #SmackDown 53.7K posts
- 15. Harden 15.3K posts
- 16. Bill Clinton 194K posts
- 17. PERTHSANTA JOY KAMUTEA 411K posts
- 18. Marjorie Taylor Greene 49.8K posts
- 19. #NEWKAMUEVENTxPerthSanta 405K posts
- 20. Bubba 59.1K posts
Something went wrong.
Something went wrong.