Commit graph

32 commits

Author SHA1 Message Date
Krrish Dholakia
29169b3039 feat(vertex_httpx.py): Moving to call vertex ai via httpx (instead of their sdk). Allows us to support all their api updates. 2024-06-12 16:47:00 -07:00
Krrish Dholakia
e391e30285 refactor: replace 'traceback.print_exc()' with logging library
allows error logs to be in json format for otel logging
2024-06-06 13:47:43 -07:00
Ishaan Jaff
f6f24b87a8 fix - lakera ai integration 2024-05-23 15:25:26 -07:00
Ishaan Jaff
60d2778d69 feat - add lakera ai prompt injection detection 2024-05-23 15:16:46 -07:00
Ishaan Jaff
f5a1274f8c feat - add debugging for moderation response 2024-05-23 13:24:09 -07:00
Ishaan Jaff
759852b1b8 feat - add open ai moderations check 2024-05-23 13:08:06 -07:00
Krrish Dholakia
aa4acaf06b fix(llm_guard.py): enable request-specific llm guard flag 2024-04-08 21:15:33 -07:00
Krrish Dholakia
c10f1d2f25 test(test_llm_guard.py): unit testing for key-level llm guard enabling 2024-03-26 17:55:53 -07:00
Krrish Dholakia
f3a56c5af2 fix(llm_guard.py): working llm-guard 'key-specific' mode 2024-03-26 17:47:20 -07:00
Krrish Dholakia
7bc76ddbc3 feat(llm_guard.py): enable key-specific llm guard check 2024-03-26 17:21:51 -07:00
Ishaan Jaff
f0992c2dbd (fix) stop using f strings with logger 2024-03-25 10:47:18 -07:00
Krrish Dholakia
0ae4906701 fix(llm_guard.py): await moderation check 2024-03-21 16:55:28 -07:00
Krrish Dholakia
860ed18a2e fix(llm_guard.py): more logging for llm guard.py 2024-03-21 11:22:52 -07:00
Krrish Dholakia
dec78ee7e5 fix: fix linting issue 2024-03-21 08:05:47 -07:00
Krrish Dholakia
e9cc6b4cc9 feat(proxy_server.py): enable llm api based prompt injection checks
run user calls through an llm api to check for prompt injection attacks. This happens in parallel to th
e actual llm call using `async_moderation_hook`
2024-03-20 22:43:42 -07:00
Krrish Dholakia
e9ff51aa70 fix(prompt_injection_detection.py): ensure combinations are actual phrases, not just 1-2 words
reduces misflagging

https://github.com/BerriAI/litellm/issues/2601
2024-03-20 19:09:38 -07:00
Krrish Dholakia
be33918438 fix(blocked_user_list.py): check if end user blocked in db 2024-03-16 13:03:52 -07:00
Krish Dholakia
42b3577fd0 Merge pull request #2498 from BerriAI/litellm_prompt_injection_detection
feat(prompt_injection_detection.py): support simple heuristic similarity check for prompt injection attacks
2024-03-13 12:28:19 -07:00
Krrish Dholakia
2f08fd55ef fix(proxy_server.py): support checking openai user param 2024-03-13 12:00:27 -07:00
Krrish Dholakia
5769bd22c3 feat(prompt_injection_detection.py): support simple heuristic similarity check for prompt injection attacks 2024-03-13 10:32:21 -07:00
Krrish Dholakia
d00773c2b1 feat(proxy_server.py): enable admin to set banned keywords on proxy 2024-02-22 18:30:42 -08:00
Krrish Dholakia
d1dd8854c2 feat(proxy_server.py): add support for blocked user lists (enterprise-only) 2024-02-22 17:51:31 -08:00
Krrish Dholakia
7089b13632 fix(llm_guard.py): add streaming hook for moderation calls 2024-02-20 20:31:32 -08:00
Krrish Dholakia
c7e7d508cb docs(enterprise.md): add llm guard to docs 2024-02-19 21:05:01 -08:00
Krrish Dholakia
66e4abcb0d feat(llm_guard.py): support llm guard for content moderation
https://github.com/BerriAI/litellm/issues/2056
2024-02-19 20:51:25 -08:00
Krrish Dholakia
6436dd41b6 docs(enterprise.md): add google text moderations to the docs 2024-02-19 14:17:52 -08:00
Krrish Dholakia
f45fd5db3b fix(llama_guard.py): fix merge 2024-02-17 23:38:53 -08:00
Krish Dholakia
f485e778cb Merge branch 'main' into litellm_google_text_moderation 2024-02-17 22:10:26 -08:00
Krish Dholakia
8e0040c09d Merge branch 'main' into litellm_llamaguard_custom_categories 2024-02-17 21:36:40 -08:00
Krrish Dholakia
ea2632d9f3 feat(google_text_moderation.py): allow user to use google text moderation for content mod on proxy 2024-02-17 18:36:29 -08:00
Krrish Dholakia
423bd4012d refactor: rename enterprise specific hooks and callbacks to be more precise 2024-02-17 17:52:45 -08:00
Krrish Dholakia
f52b3c5f84 feat(llama_guard.py): allow user to define custom unsafe content categories 2024-02-17 17:42:47 -08:00