 |
Work Item Reference | ETSI Doc. Number |
STF | Technical Body in Charge |
Standard Not Ready For Download |
||
| DGS/ENI-0048v411_LLMAttack | GS ENI 048 | ENI | |||||
| Current Status (Click to View Full Schedule) |
Latest Version |
Cover Date | Standstill | Creation Date | |||
|
|
View Standstill Information | 2024-06-12 | |||||
| Rapporteur | Technical Officer | Harmonised Standard | |||||
Ziting Zhang
|
Christine Mera
|
No | |||||
| Title |
Experiential Networked Intelligence (ENI); Content security testing of large language model (LLM) against prompt injection attacks |
||||||
|
Scope and Field of Application |
This work item applies to the content security testing of large language models against prompt injection attacks. This WI will specify: 1. Requirements of content security testing of LLM against prompt injection attacks. 2. Definition of the test objects, type of prompt injection attacks on LLM. 3. A standard and comprehensive procedures to test and evaluate the content security capacity of the under-test large language model when facing prompt injection attacks, including the construction of test samples, evaluation metrics and evaluation method, etc. |
||||||
|
Supporting Organizations |
China Telecom AsiaInfo CAICT Huawei UK ZTE Futurewei | ||||||
