microsoft / microsoft/MInference
[Question]: vertical slash pattern
Open
@iofu728 is already working on this.
Since Jul 7, 2024.
question
- Dominant language
- Python
- Stars
- 1.2k
- Forks
- 82
- Avg merge
- 1d 18h
- Merged PRs (30d)
- 1
Description
Describe the issue
Hello,
Vertical lines in attention correspond to "heavy hitters", tokens that are attended every time.
I don't really get what the intuition behind the off diagonal lines in the attention matrix. Is it only an "artifact" related to the periodicity of RoPE ? What is your understanding of this phenomenon ?
Contributor guide
No contributing guide indexed for this repository
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Assessment
This issue has not been assessed yet.