Anthropic Warns That Minimal Data Contamination Can ‘Poison’ Large AI Models

Anthropic has warned that even a few poisoned samples in a dataset can compromise an AI model. A joint study with the UK AI Security Institute found that as few as 250 malicious documents can implant backdoors in LLMs up to 13B parameters, proving model size offers no protection.

from Gadgets 360 https://ift.tt/pCQ2ReA

Comments

Popular posts from this blog

Samsung Galaxy S26 Ultra Tipped to Get Under-Display Camera (UDC) Similar to Galaxy Z Fold 6

FBI Accuses North Korean Hackers of Bybit Attack, Urges Crypto Firms to Block Linked Transactions

Jio Choice Number: What is it and How to Buy Jio Fancy Number Using Different Methods