github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-07-17 05:36:16 -05:00
Code Issues 16.9k Packages Projects Releases 122 Wiki Activity

[GH-ISSUE #2820] Ollama behaves weird - tries to force number outputs on nearly every prompt #1712

New Issue
Open
opened 2026-04-12 11:41:15 -05:00 by GiteaMirror · 0 comments
GiteaMirror commented 2026-04-12 11:41:15 -05:00
Owner
Copy Link

Originally created by @K0IN on GitHub (Feb 29, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/2820

Im buildng a chrome extension that uses ollama.js to predict input.
I noticed that i get number suggestions way more often than, coherent text.

Here is my code that calls the LLM (note that the input is just the text field value)

image

and here is an example recording (you can see the full prompt + output on the right) -> this is generated by the code above:

test (1).webm

I tested the following models, all behave like that:

mistral:7b-text, llama2:text, gemma:7b-text, gemma:7b-text-fp16

Here is the command i used for ollam to run

OLLAMA_ORIGINS=chrome-extension://* ollama serve

with the latest Version of ollama (installed on wsl with GPU support)

My assumption is that text models continue the text in a coherent way, so I can just drop in the text input and it will continue my sentence.

Originally created by @K0IN on GitHub (Feb 29, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/2820 Im buildng a chrome extension that uses ollama.js to predict input. I noticed that i get number suggestions way more often than, coherent text. Here is my code that calls the LLM (note that the input is just the text field value) ![image](https://github.com/ollama/ollama/assets/19688162/903edb3d-10da-40f6-ba0b-d12988bc2e5e) and here is an example recording (you can see the full prompt + output on the right) -> this is generated by the code above: [test (1).webm](https://github.com/ollama/ollama/assets/19688162/213d9d73-79f6-43cd-a8f8-7e98f020e0c7) I tested the following models, all behave like that: mistral:7b-text, llama2:text, gemma:7b-text, gemma:7b-text-fp16 Here is the command i used for ollam to run > OLLAMA_ORIGINS=chrome-extension://* ollama serve with the latest Version of ollama (installed on wsl with GPU support) My assumption is that text models continue the text in a coherent way, so I can just drop in the text input and it will continue my sentence.
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
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#1712
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?