CyberRota Analysis
AI-GeneratedThe vulnerability affects the vLLM inference engine prior to version 0.24.0, where a specific multi-request workload can lead to a denial of service by causing the engine to crash due to an out-of-vocabulary token being processed. This issue can be exploited remotely via public gRPC endpoints, impacting all clients using the shared engine worker. Organizations utilizing vLLM for serving large language models should prioritize upgrading to version 0.24.0 to mitigate the risk of service disruption.
Public Exploit Signal
A public exploit, PoC, GitHub repository or Metasploit reference was detected for this CVE.
Note: these links are listed for security research and verification purposes only.
Original NVD Description
vLLM is a high-throughput and memory-efficient inference and serving engine for LLMs. Prior to 0.24.0, a frontend-legal multi-request speculative decoding workload can cause the rejection sampler to produce a recovered token equal to the model vocabulary size boundary value, which is then converted to negative one when the engine selects the next live token for a request and is written back into the drafter's input ids; that out-of-vocabulary value is later consumed by the model's embedding and attention path and crashes the engine worker with a GPU device-side assertion. The same triggering request sequence is reachable through the public gRPC Generate and Abort endpoints, so a remote client that can send generation requests can crash the shared engine worker, aborting concurrent requests and causing a service-wide denial of service for other clients of the deployment until the worker is restarted. This issue is fixed in version 0.24.0.
Related CVEs
Other vulnerabilities affecting the same vendor(s)