LLM10 typos

[mbomb007]

Top 10 for LLMs

Example 7 in the "Common Examples of Vulnerability" sub-section mentions Example 5, but I think "Example 6" would be the correct reference, as number 5 was a side-channel attack.

This bypasses the limitations of traditional query-based extraction used in Example 5

There is also this typo:

could infiltrate from insider or outside

Should be "inside" not "insider".

[GangGreenTemperTatum]

thanks @mbomb007 for reporting this! since it's the PDF in question, this won't most likely be fixed until v2.0 publishing but it's most likely somewhere in the site and repo too potentially, i'll triage this to design leads

[GangGreenTemperTatum]

@mbomb007

Example 7 in the "Common Examples of Vulnerability" sub-section mentions Example 5, but I think "Example 6" would be the correct reference, as number 5 was a side-channel attack.

i don't see any Example 7 in either of the versions:

https://github.com/OWASP/www-project-top-10-for-large-language-model-applications/blob/main/2_0_vulns/LLM10_ModelTheft.md https://github.com/OWASP/www-project-top-10-for-large-language-model-applications/blob/main/1_1_vulns/LLM10_ModelTheft.md

Should be "inside" not "insider".

thanks and fixed! https://github.com/OWASP/www-project-top-10-for-large-language-model-applications/pull/276

[mbomb007]

i don't see any Example 7 in either of the versions:

That's because Example 2 has already been removed in the markdown (which was the other issue I reported). In the PDF, it hasn't, so examples 5 and 6 are instead 6 and 7. It's correct in the markdown.

[GangGreenTemperTatum]

@rossja 🆙