Who's Harry Potter? Approximate Unlearning in LLMs
Paper
•
2310.02238
•
Published
•
4
Ronen Eldan
roneneldan
AI & ML interests
None yet
Recent Activity
authored
a paper
8 days ago
Phi-4 Technical Report
Organizations
roneneldan's activity
add some metadata
#20 opened 4 months ago
by
davanstrien
Better formatting of hyperparams and code snippet
#19 opened 5 months ago
by
b-gendron
commented
a
paper
11 months ago
Number of parameters in the model
3
#16 opened 12 months ago
by
bivouac0
Correct maximum positional embeddings
3
#17 opened 11 months ago
by
nixgd
How is this dataset encoded?
2
#10 opened about 1 year ago
by
Tategami
Instruction-following variant of `TinyStoriesV2-GPT4-*.txt`
2
#11 opened about 1 year ago
by
ivnle
Context Length of the model 512 or 2048?
2
#12 opened about 1 year ago
by
maveriq
MetaData for the GPT4 train/test files
1
#9 opened over 1 year ago
by
vgoklani
Can we get the extended dataset please?
5
#5 opened over 1 year ago
by
dnhkng
validation scores
1
#10 opened over 1 year ago
by
AntonioMartini
Unable to replicate generations from paper
1
#9 opened over 1 year ago
by
solver1104
how does one "keep the top 10K most common tokens"?
4
#4 opened over 1 year ago
by
nawal2
How long was this model trained?
2
#8 opened over 1 year ago
by
jploski
Tinystories in other languages
3
#7 opened over 1 year ago
by
mukundesh
Model training schedule and parameters
3
#3 opened over 1 year ago
by
AntonioMartini
Training data
1
#4 opened over 1 year ago
by
AntonioMartini
broken dataset - unicode error
13
#1 opened over 1 year ago
by
JHare
row structure of the dataset?
2
#2 opened over 1 year ago
by
infiniteperplexity
what is the license for this model?
1
#5 opened over 1 year ago
by
omeze