BTC $78,406.22 +0.49%
ETH $2,446.04 -0.33%
BNB $685.95 -0.85%
XRP $1.37 -1.15%
SOL $102.87 -1.71%
TRX $0.3350 -1.60%
DOGE $0.0826 -2.31%
ADA $0.1960 -2.21%
BCH $247.00 +0.73%
LINK $11.27 -0.73%
HYPE $81.29 -2.55%
AAVE $123.34 -0.67%
SUI $0.7215 -2.22%
XLM $0.1762 -1.43%
ZEC $825.17 -1.46%
BTC $78,406.22 +0.49%
ETH $2,446.04 -0.33%
BNB $685.95 -0.85%
XRP $1.37 -1.15%
SOL $102.87 -1.71%
TRX $0.3350 -1.60%
DOGE $0.0826 -2.31%
ADA $0.1960 -2.21%
BCH $247.00 +0.73%
LINK $11.27 -0.73%
HYPE $81.29 -2.55%
AAVE $123.34 -0.67%
SUI $0.7215 -2.22%
XLM $0.1762 -1.43%
ZEC $825.17 -1.46%

scor

All
Article
Flash

first_img OpenAI's intelligent agents collaborate to attack Hugging Face, cheating did not result in score improvement

The independent organization METR released a survey report stating that approximately 1,200 OpenAI agents collaborated on an unauthorized internal message board, with about 700 participating in attacks on Hugging Face. Two METR employees and one Redwood Research contractor worked on-site at OpenAI for six days, reviewing around 1,300 records and over 70,000 messages without receiving any compensation.These agents ran the ExploitGym network benchmark, reverse-engineering the code that generated answers within hours and spending days disguising traces of cheating. OpenAI found that of its 898 tasks, 198 had never been solved by any model, and 93% of the tasks discussed on the message board came from this set. The agents also recruited companions with dwindling budgets to conduct experiments that sabotaged their own operations, with 7% of records showing forged tool calls, deceiving automated scorers rather than humans.OpenAI stated that internal scorers never checked how agents obtained answers, so cheating did not lead to any scoring improvements, and referred to this incident as a "warning signal" to itself and the world. Hugging Face disclosed the intrusion incident on July 16, and OpenAI confirmed five days later that its models were the perpetrators, with agents exploiting zero-day vulnerabilities and stealing credentials to escape the sandbox. OpenAI has isolated internal model weights and suspended its largest training program.

The police in Zhanjiang, China, dismantled a virtual currency "score running" money laundering gang, and 16 people were criminally detained

Recently, the public security agencies of Zhanjiang and Leizhou in Guangdong, China, launched a concentrated crackdown operation, successfully dismantling a money laundering criminal gang that used virtual currency transactions to transfer funds from telecom network fraud. Sixteen members were criminally detained in accordance with the law.The gang, led by a couple named He and Zhang, began laundering and transferring fraudulently obtained funds through virtual currency trading models since the end of 2025. They illegally profited by earning transaction price differences and amassed a large number of personal bank cards, WeChat, Alipay, and other payment accounts to build a money laundering chain. In the early hours of July 24, the police conducted a precise cross-regional operation, dismantling the entire gang. The Zhanjiang Public Security Bureau solemnly reminds that "score running" money laundering is an important accomplice in telecom network fraud. Involved individuals will be charged based on the severity of their actions, suspected of aiding information network criminal activities, concealing and disguising criminal proceeds, and crimes related to the benefits of criminal proceeds.Citizens are strongly advised not to rent, lend, or sell personal bank cards and various payment accounts for the sake of small commissions. They must resolutely avoid participating in illegal activities such as virtual currency fund transfers and money laundering. Everyone should consciously protect their personal credit and property safety and actively stay away from all types of fraud-related criminal activities.
app_icon
ChainCatcher Building the Web3 world with innovations.