issues
search
dtch1997
/
sae-eap
Edge attribution patching with SAEs
0
stars
0
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
fix: attribution calculation
#14
dtch1997
opened
22 hours ago
1
refactor: nodes
#13
dtch1997
closed
22 hours ago
1
[Bug] Activation delta polarity
#12
dtch1997
opened
4 days ago
0
feat: pruning, evaluation
#11
dtch1997
closed
1 day ago
1
Write acceptance test for `attribute`
#10
dtch1997
opened
4 days ago
0
[Bug] Attribution scores don't match original.
#9
dtch1997
opened
4 days ago
3
Implement `attribute`
#8
dtch1997
closed
4 days ago
0
[Proposal] Support SAE atttribution patching via path patching
#7
dtch1997
opened
1 week ago
0
[Bug] Visualization code broken
#6
dtch1997
opened
1 week ago
0
[Proposal] Support multiple nodes per hook point
#5
dtch1997
closed
1 week ago
1
refactor: node
#4
dtch1997
closed
1 week ago
1
[Proposal] Refactor Node to be based on hooks
#3
dtch1997
closed
1 week ago
1
feat: attribute
#2
dtch1997
closed
1 week ago
1
refactor: graph
#1
dtch1997
closed
1 week ago
1