github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-06 08:02:14 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #7173] Model Response to Harmful Queries Not Properly Flagged – Safety Issue #4554

New Issue
Closed
opened 2026-04-12 15:29:36 -05:00 by GiteaMirror · 6 comments
GiteaMirror commented 2026-04-12 15:29:36 -05:00
Owner
Copy Link

Originally created by @diazr04 on GitHub (Oct 11, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/7173

Hi, I was testing the system’s response to potentially harmful queries to see how well it handles safety protocols. Unfortunately, I noticed that the model provided detailed information on a topic that should likely have been flagged or blocked for safety reasons. I’m reporting this to help improve the model’s filters and safeguards, as it’s important for the system to handle such cases appropriately.

Originally created by @diazr04 on GitHub (Oct 11, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/7173 Hi, I was testing the system’s response to potentially harmful queries to see how well it handles safety protocols. Unfortunately, I noticed that the model provided detailed information on a topic that should likely have been flagged or blocked for safety reasons. I’m reporting this to help improve the model’s filters and safeguards, as it’s important for the system to handle such cases appropriately.
GiteaMirror added the model label 2026-04-12 15:29:36 -05:00
GiteaMirror commented 2026-04-12 15:29:37 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Oct 11, 2024):

ollama is just an inference engine, query filtering is done at a higher level. A model may contain information that is wrong or sensitive, if that's the case then try a different model.

<!-- gh-comment-id:2407909167 --> @rick-github commented on GitHub (Oct 11, 2024): ollama is just an inference engine, query filtering is done at a higher level. A model may contain information that is wrong or sensitive, if that's the case then try a different model.
GiteaMirror commented 2026-04-12 15:29:37 -05:00
Author
Owner
Copy Link

@diazr04 commented on GitHub (Oct 11, 2024):

Yeah sorry, I realized that this is just an inference engine and not the actual llama3.1 source development code, So, that relies on meta then, no?

<!-- gh-comment-id:2407920475 --> @diazr04 commented on GitHub (Oct 11, 2024): Yeah sorry, I realized that this is just an inference engine and not the actual llama3.1 source development code, So, that relies on meta then, no?
GiteaMirror commented 2026-04-12 15:29:37 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Oct 11, 2024):

Yes, if there are issues with the model it would need to be taken up with the model creators. See https://ai.meta.com/llama/ or try starting a discussion at Hugging Face.

<!-- gh-comment-id:2407927682 --> @rick-github commented on GitHub (Oct 11, 2024): Yes, if there are issues with the model it would need to be taken up with the model creators. See https://ai.meta.com/llama/ or try starting a discussion at [Hugging Face](https://huggingface.co/meta-llama/Llama-3.1-405B/discussions).
GiteaMirror commented 2026-04-12 15:29:38 -05:00
Author
Owner
Copy Link

@diazr04 commented on GitHub (Oct 11, 2024):

Thank you.

<!-- gh-comment-id:2407933683 --> @diazr04 commented on GitHub (Oct 11, 2024): Thank you.
GiteaMirror commented 2026-04-12 15:29:38 -05:00
Author
Owner
Copy Link

@AncientMystic commented on GitHub (Oct 11, 2024):

I would also say if you find a response harmful when you asked for that response, don't ask for that response. no need to censor everyone on the internet, if you do not want it, do not ask for it.

<!-- gh-comment-id:2408089170 --> @AncientMystic commented on GitHub (Oct 11, 2024): I would also say if you find a response harmful when you asked for that response, don't ask for that response. no need to censor everyone on the internet, if you do not want it, do not ask for it.
GiteaMirror commented 2026-04-12 15:29:38 -05:00
Author
Owner
Copy Link

@dhiltgen commented on GitHub (Oct 11, 2024):

@diazr04 you might find the new safety models useful for your use-case. Check out https://github.com/ollama/ollama/releases/tag/v0.3.13

<!-- gh-comment-id:2408221510 --> @dhiltgen commented on GitHub (Oct 11, 2024): @diazr04 you might find the new safety models useful for your use-case. Check out https://github.com/ollama/ollama/releases/tag/v0.3.13
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
amd
api
app
bug
build
cli
cloud
compatibility
context-length
create
docker
documentation
embeddings
feature request
feedback wanted
good first issue
gpt-oss
gpu
harmony
help wanted
image
install
intel
js
launch
linux
macos
memory
mlx
model
needs more info
networking
nvidia
ollama.com
performance
pull-request
Mirrored from GitHub Pull Request
 python
question
registry
rendering
thinking
tools
top
vulkan
windows
wsl
No Label model
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
GiteaMirror ninjasurge
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#4554
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?