Researchers find LLMs like ChatGPT output sensitive data even after it’s been ‘deleted’
According to the scientists, there’s no universal method by which data can be deleted from a pretrained large language model.
Archive context
Older archive item. Useful for background and entity history, but not a fresh market-moving signal.
According to the scientists, there’s no universal method by which data can be deleted from a pretrained large language model.
Why this matters
This research story adds another data point to the current market tape and is useful when read alongside nearby source coverage.
Original source
Read on CointelegraphRelated market context
Bitcoin Price Prediction: Saylor’s Strategy is a Risk to Bitcoin, According to JP Morgan
JPMorgan has flagged a structural risk most Bitcoin price prediction bulls haven’t priced in: the same entity driving the most agg...
Vitalik Buterin Says Ethereum’s Next Major Upgrade Will Take Three to Four Years and Rival the Merge
Ethereum co-founder Vitalik Buterin says the network’s next major overhaul, an initiative researchers are calling “Lean Ethereum,”...
OPEC+ raises crude output despite falling prices, and crypto markets are paying attention
OPEC+'s output increase amid falling prices could ease inflation, potentially leading to lower interest rates and boosting risk as...
CryptoQuant Flags Exchange Deposit Spike As Bitcoin Volatility Risk Builds
Bitcoin’s rebound has not removed the risk of another volatile move. CryptoQuant is warning that exchange deposit activity has pic...
Bitcoin Rebounds Toward $63,000, But ETF Flows Still Hold The Key
Bitcoin has started the new week with a little more colour on the screen. After sliding through the kind of levels that usually tr...
Brazil vs Norway at the 2026 World Cup is also a crypto market event
The integration of crypto in global sports events like the World Cup highlights its growing influence on fan engagement and market...