Pietro Lesci
pietrolesci
AI & ML interests
I like developing and applying causal methods to study the effect of training choices on models’ behaviour, including memorisation, shortcut learning, and tokenisation.
Organizations
Create README.md
#2 opened 9 months ago
by pietrolesci
Domain and provenance annotation
9
#1 opened over 2 years ago
by 
haukur
Trapezoidal scheduler with cooldown phase
👍 1
3
#4 opened over 1 year ago
by 
maveriq
Tokenizer `merges.txt` files
3
#5 opened almost 2 years ago
by 
pietrolesci
Sequence "packing" logic
👍 2
2
#2 opened about 2 years ago
by pietrolesci
Pad-only sequences from mmap'ed dataset after a certain index
#1 opened over 2 years ago
by pietrolesci
Add full sequences (beyond the first 64 tokens)
3
#1 opened over 2 years ago
by pietrolesci
Add full sequences (beyond the first 64 tokens)
3
#1 opened over 2 years ago
by pietrolesci
Domain and provenance annotation
9
#1 opened over 2 years ago
by haukur