github/langchain
1
0
Fork 0
mirror of https://github.com/hwchase17/langchain.git synced 2025-09-24 03:52:10 +00:00
Code Issues Packages Projects Releases Wiki Activity

experimental[patch]: Fix import errors and docs of experimental prompt injection identifier (#21586)

Browse Source 
**Description**

I noticed that code in experimental was not functional due to small
errors, and users were complaining at
https://stackoverflow.com/questions/78272521/hugging-face-prompt-injection-identifier
and https://github.com/langchain-ai/langchain/discussions/19995.

I know folks are busy with `v2.0` now, but I believe this is worthy fix
because it is isolated, minor, backwards compatible, and fixes an error
with current v0.1 code

**Issue**
https://github.com/langchain-ai/langchain/discussions/19995

**Dependencies**
None

**Twitter handle**
https://twitter.com/nelsonauner


cc @asofter

---------

Co-authored-by: Bagatur <baskaryan@gmail.com>
This commit is contained in:
Nelson Auner
2024-05-21 16:29:25 -07:00
committed by GitHub
parent 5a40413bfd
commit 946a0a38bd
2 changed files with 6 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
docs/docs/guides/productionization/safety/hugging_face_prompt_injection.ipynb
View File

@@ -50,7 +50,7 @@
"from transformers import AutoTokenizer, pipeline\n",
"\n",
"# Using https://huggingface.co/protectai/deberta-v3-base-prompt-injection-v2\n",
"model_path = \"laiyer/deberta-v3-base-prompt-injection-v2\"\n",
"model_path = \"protectai/deberta-v3-base-prompt-injection-v2\"\n",
"revision = None # We recommend specifiying the revision to avoid breaking changes or supply chain attacks\n",
"tokenizer = AutoTokenizer.from_pretrained(\n",
" model_path, revision=revision, model_input_names=[\"input_ids\", \"attention_mask\"]\n",