The Pile Companion
Collection
4 items
•
Updated
Pietro Lesci PRO
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.
Recent Activity
updated
a collection
8 days ago
The Pile Companion
updated
a collection
8 days ago
The Pile Companion
updated
a dataset
8 days ago
InfoTokenizers/finewebedu-20B
Organizations
pietrolesci's activity
updated
a
collection
8 days ago
updated
a
dataset
8 days ago
updated
a
dataset
14 days ago
updated
10
datasets
20 days ago
pietrolesci/pythia-deduped-memorisation-profiles
Viewer
•
Updated
•
2.13M
•
46
pietrolesci/pile-validation
Viewer
•
Updated
•
429k
•
176
pietrolesci/pile-deduped-subset
Viewer
•
Updated
•
16.3k
•
25
pietrolesci/pythia-deduped-stats
Viewer
•
Updated
•
16.3M
•
1.19k
pietrolesci/pythia-deduped-stats-raw
Viewer
•
Updated
•
14.9M
•
20.4k
pietrolesci/agnews
Viewer
•
Updated
•
510k
•
103
pietrolesci/amazoncat-13k
Viewer
•
Updated
•
5.99M
•
421
•
1
pietrolesci/wikitoxic
Viewer
•
Updated
•
894k
•
153
•
1
pietrolesci/multiwoz_all_versions
Viewer
•
Updated
•
82k
•
79
•
1
pietrolesci/anchoral-paper-artefacts
Viewer
•
Updated
•
2.78M
•
4.7k
authored
a
paper
20 days ago
published
a
dataset
26 days ago
updated
a
model
27 days ago
published
a
model
27 days ago
updated
a
collection
about 1 month ago