Explore tweets tagged as #nullifAI
#threatreport #MediumCompleteness Malicious ML models discovered on Hugging Face platform | 06-02-2025 Source: https://t.co/MMiLeUyWQz Key details below ↓ 💀Threats: Nullifai_technique, Supply_chain_technique, 🎯Victims: Hugging face 🏭Industry: Software_development 🌐Geo:
0
0
0
🤖 Learn about nullifAI & other AI-related threats to software supply chains in our next webinar on 2/20 at 11am ET: https://t.co/Ev1NADqPOH
2
0
3
#threatreport #LowCompleteness Malicious attack method on hosted ML models now targets PyPI | 23-05-2025 Source: https://t.co/cbvkwF0EtA Key details below ↓ 💀Threats: Supply_chain_technique, Nullifai_technique, 🎯Victims: Developers, Users of alibaba ai labs services
0
0
0
Researchers from ReversingLabs recently discovered two Hugging Face models containing malicious code. The nullifAI attack involves abusing Pickle file serialization. https://t.co/LlfGYQ6zFh
0
11
56
Interesting report from ReversingLabs researchers, who named a new attack nullifAI, a novel malware distribution technique targeting ML models on Hugging Face. 😈Attackers exploited Pickle serialization to deliver payloads undetected. https://t.co/wwGP91Bpcg
3
39
103
Malicious AI models on Hugging Face exploit a novel attack technique called nulliFai. Learn how they bypass protective measures: https://t.co/y810RQ9QuH
0
0
0
Thanks AK for sharing our work! Unlike autoregressive LLMs, diffusion LLMs can be conditioned on future reasoning hints during generation through inpainting 🧩, enabling guided exploration toward correct solutions. We show that applying inpainting-guided exploration in RL
0
0
0
⚠️ #ML devs, take note: RL threat researchers have identified nullifAI, a novel attack technique used on ML models hosted on #HuggingFace.
1
4
4
Malicious ML Models on Hugging Face Exploit Broken Pickle Format Cybersecurity researchers have found two malicious machine learning models on Hugging Face. These models use "broken" pickle files to evade detection, a technique called "nullifAI". The payload is a platform-aware
0
0
0
New nullifAI technique bypasses Hugging Face protective measures, raising cybersecurity concerns. Learn more:
0
0
0
オープンソースの AI モデル: 悪意のあるコードや脆弱性のパーフェクトストーム Open Source AI Models: Perfect Storm for Malicious Code, Vulnerabilities #DarkReading (Feb 15) #オープンソース #AIモデル #セキュリティリスク #HuggingFace #NullifAI
https://t.co/qBlGeDvbhk
0
0
2
Malicious ML models discovered on Hugging Face platform Software development teams working on machine learning take note: RL threat researchers have identified nullifAI, a novel attack technique used on Hugging Face. https://t.co/XrxESmqlhP
0
0
1
Hugging Face｜AIモデルの新たな脆弱性「nullifAI」が発見 – 破損Pickleファイルを悪用した検知回避手法とは https://t.co/FQB6Ibc6Wf
0
0
0
1/7 🚨 Malicious ML models on Hugging Face exploit "broken" pickles to evade detection! 📛 nullifAI uses reverse shells to connect to hardcoded IPs, threatening supply chain security. 🔒 #Cybersecurity #MLSecurity #HuggingFace 🚀
0
0
0
Hugging Face｜AIモデルの新たな脆弱性「nullifAI」が発見 – 破損Pickleファイルを悪用した検知回避手法とは https://t.co/8Efs2WhDlw
0
1
1
Hugging Face｜AIモデルの新たな脆弱性「nullifAI」が発見 – 破損Pickleファイルを悪用した検知回避手法とは - innovaTopia
1
0
2