github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-07 16:40:08 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #6706] Reflection 70B has significant issue with the weights #66259

New Issue
Closed
opened 2026-05-04 01:32:47 -05:00 by GiteaMirror · 4 comments
GiteaMirror commented 2026-05-04 01:32:47 -05:00
Owner
Copy Link

Originally created by @gileneusz on GitHub (Sep 9, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/6706

The whole drama is described here:

https://x.com/shinboson/status/1832933753837982024

sorry for recommending the model, I was unaware of that and easily got into the hype

it's still possible that's just technical issue, but I'm suspicious:

https://x.com/rohanpaul_ai/status/1833094994929897862/photo/1

Originally created by @gileneusz on GitHub (Sep 9, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/6706 The whole drama is described here: https://x.com/shinboson/status/1832933753837982024 sorry for recommending the model, I was unaware of that and easily got into the hype it's still possible that's just technical issue, but I'm suspicious: https://x.com/rohanpaul_ai/status/1833094994929897862/photo/1
GiteaMirror added the model label 2026-05-04 01:32:47 -05:00
GiteaMirror commented 2026-05-04 01:33:02 -05:00
Author
Owner
Copy Link

@nisten commented on GitHub (Sep 9, 2024):

You have to fix the damn tokenizer_config.json before it works correctly on llama.cpp , please stop making assumptions on twitter threads about a broken 3rd party api , i was not able to reproduce anything the twitter thread said when running it locally.
Screenshot 2024-09-09 at 12 51 43 AM

<!-- gh-comment-id:2337964298 --> @nisten commented on GitHub (Sep 9, 2024): You have to fix the damn tokenizer_config.json before it works correctly on llama.cpp , please stop making assumptions on twitter threads about a broken 3rd party api , i was not able to reproduce anything the twitter thread said when running it locally. ![Screenshot 2024-09-09 at 12 51 43 AM](https://github.com/user-attachments/assets/6a8c5e20-20b6-49d9-ae6c-8e8610a45f87)
GiteaMirror commented 2026-05-04 01:33:06 -05:00
Author
Owner
Copy Link

@juanjgit commented on GitHub (Sep 9, 2024):

I'm 99% sure this is a scam.

It's not just "Twitter threads." Multiple independent researchers from the LLM community have analyzed both the API and the local model.

Yes, there's a 1% chance that a single individual who releases an LLM he admits to developing in only 15 days using a known technique has found the holy grail by creating a 70B model that surpasses models from OpenAI and Claude. Yes, there's a 1% chance he used a fake API, but only by mistake. Yes, there's a 1% chance of an unknown and unexplained "upload error" causing a bug that makes the model seem less capable.

But the fact is, no one has been able to reproduce his metrics. There's no single piece of concrete proof that the model is as intelligent as he claims. It's highly likely a scam. It should be removed from Ollama.

<!-- gh-comment-id:2338827702 --> @juanjgit commented on GitHub (Sep 9, 2024): I'm 99% sure this is a scam. It's not just "Twitter threads." Multiple independent researchers from the LLM community have analyzed both the API and the local model. Yes, there's a 1% chance that a single individual who releases an LLM he admits to developing in only 15 days using a known technique has found the holy grail by creating a 70B model that surpasses models from OpenAI and Claude. Yes, there's a 1% chance he used a fake API, but only by mistake. Yes, there's a 1% chance of an unknown and unexplained "upload error" causing a bug that makes the model seem less capable. But the fact is, no one has been able to reproduce his metrics. There's no single piece of concrete proof that the model is as intelligent as he claims. It's highly likely a scam. It should be removed from Ollama.
GiteaMirror commented 2026-05-04 01:33:10 -05:00
Author
Owner
Copy Link

@kowshik24 commented on GitHub (Sep 9, 2024):

I have tried just a simple question! 😉

image

<!-- gh-comment-id:2338969379 --> @kowshik24 commented on GitHub (Sep 9, 2024): I have tried just a simple question! 😉 ![image](https://github.com/user-attachments/assets/b9428742-c40b-4de3-8c87-4b9520018343)
GiteaMirror commented 2026-05-04 01:33:13 -05:00
Author
Owner
Copy Link

@jmorganca commented on GitHub (Sep 12, 2024):

https://ollama.com/library/reflection is updated to the latest available weights: https://huggingface.co/mattshumer/ref_70_e3

Keep in mind the controversy around this model...

<!-- gh-comment-id:2345065708 --> @jmorganca commented on GitHub (Sep 12, 2024): https://ollama.com/library/reflection is updated to the latest available weights: https://huggingface.co/mattshumer/ref_70_e3 Keep in mind the controversy around this model...
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#66259
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?