8 matches found
Activation Surgery: Jailbreaking White-Box LLMs without Touching the Prompt
Most jailbreak techniques for Large Language Models LLMs primarily rely on prompt modifications, including paraphrasing, obfuscation, or conversational strategies. Meanwhile, abliteration techniques also known as targeted ablations of internal components have been used to study and explain LLM...
Integer Overflow lead to DOS in handling Accept-Encoding header in API /v2/models/<model-name>/generate
This report is not public...
EUVD-2019-12207
Malware in sbrugna...
Optimized Couplings for Watermarking Large Language Models
Large-language models LLMs are now able to produce text that is, in many cases, seemingly indistinguishable from human-generated content. This has fueled the development of watermarks that imprint a signal'' in LLM-generated text with minimal perturbation of an LLM's output. This paper provides a...
Modelina 代码注入漏洞
Modelina is the asyncapi personal developer's library for generating data models based on input such as AsyncAPI, OpenAPI or JSON schema documents. A code injection vulnerability exists in versions of Modelina prior to 1.0.0 that stems from vulnerability to code injection attacks...
CVE-2019-2567
Vulnerability in the Oracle Configurator component of Oracle Supply Chain Products Suite subcomponent: Active Model Generation. Supported versions that are affected are 12.1 and 12.2. Easily exploitable vulnerability allows unauthenticated attacker with network access via HTTP to compromise Oracl...
CVE-2019-2567
Vulnerability in the Oracle Configurator component of Oracle Supply Chain Products Suite subcomponent: Active Model Generation. Supported versions that are affected are 12.1 and 12.2. Easily exploitable vulnerability allows unauthenticated attacker with network access via HTTP to compromise Oracl...
CVE-2019-2567
CVE-2019-2567 affects Oracle Configurator (Active Model Generation) in Oracle Supply Chain Products Suite. Affected: 12.1 and 12.2. Issue: unauthenticated attacker over HTTP can gain unauthorized access to Oracle Configurator data. CVSSv3 base score 7.5 (High) with confidentiality impact. Connect...