-
Notifications
You must be signed in to change notification settings - Fork 8
Updating EigenAI overview #217
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
base: main
Are you sure you want to change the base?
Conversation
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
|
|
||
| :::note | ||
| - OpenAI-compatible refers to the [messages-based Chat Completions API](https://platform.openai.com/docs/api-reference/chat/create). | ||
| - Deterministic refers to providing one request (prompt, seed/parameters) to the EigenAI API multiple times will produce the same output bit-by-bit. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
| - Deterministic refers to providing one request (prompt, seed/parameters) to the EigenAI API multiple times will produce the same output bit-by-bit. | |
| - Deterministic behavior refers to providing one request (prompt, seed/parameters) to the EigenAI API multiple times and receiving the same output bit-by-bit every time. The EigenAI inference stack was designed with this determinism guarantee in mind. **This behavior is not based on caching the results.** Safety-critical systems cannot be vulnerable to potential cache misses. | |
| - A directionally similar work was published by [Thinking Machines, titled Defeating Nondeterminism in LLMs](https://thinkingmachines.ai/blog/defeating-nondeterminism-in-llm-inference/). EigenAI achieves the same outcome as this research, making EigenAI the world's first production-level deterministic LLM API. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Updated
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I broke the piece about not caching into a separate bullet point and reworded slightly - you're saying that 'determinism requires not caching the results' right? Or something different to that?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Not that it requires not caching, it's that the deterministic behavior is not a result of caching, since typically caching gives you deterministic behavior.
No description provided.