Hacker News
new
|
past
|
comments
|
ask
|
show
|
jobs
|
submit
login
d110af5ccf
on July 14, 2023
|
parent
|
context
|
favorite
| on:
Faster neural networks straight from JPEG (2018)
Information loss, or the result of useful computation? VAEs exist after all.
chriswarbo
on July 14, 2023
[–]
LLMs can't reason about spelling, e.g. asking for a sentence which contains no letter "a"; and can also struggle with rhyming, etc. The most obvious explanation is that they never 'see' the underlying letters/spelling, only tokens.
Guidelines
|
FAQ
|
Lists
|
API
|
Security
|
Legal
|
Apply to YC
|
Contact
Search: