issues
search
AlignmentResearch
/
tuned-lens
Tools for understanding how transformer predictions are built layer-by-layer
https://tuned-lens.readthedocs.io/en/latest/
MIT License
432
stars
47
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
KKwargs don't propogate to `load_lens_artifacts` and `th.load`
#85
levmckinney
closed
1 year ago
0
Fix kwarg propagation in `TunedLens.from_model_and_pretrained()`
#84
levmckinney
closed
1 year ago
1
Figure out why tests sometimes hang in pre-merge checks
#83
levmckinney
closed
1 year ago
0
Stricter linting and type checking in pre-commit
#82
levmckinney
closed
1 year ago
1
Implementation RMSNorm to support LLaMa
#81
levmckinney
closed
1 year ago
1
Code purge
#80
levmckinney
closed
1 year ago
3
Check pointing and resuming from checkpoints
#79
levmckinney
closed
1 year ago
2
Refactors and adds tests for`tuned_lens.data`
#78
levmckinney
closed
1 year ago
0
WIP - Resolves #73 - Implement first order differences of prediction trajectories
#77
alexmlong
closed
1 year ago
0
WIP - Resolves #73 - Implement first order differences of prediction trajectories
#76
alexmlong
closed
1 year ago
0
Scripts cleanup
#75
levmckinney
closed
1 year ago
1
Fixed and simplified .codecov.yaml
#74
levmckinney
closed
1 year ago
1
Implement first order differences of prediction trajectories
#73
levmckinney
opened
1 year ago
0
Prediction depth
#72
levmckinney
opened
1 year ago
2
Removing assumptions from the `Decoder`
#71
levmckinney
closed
1 year ago
0
Support RMSNorm
#70
levmckinney
closed
1 year ago
1
No heuristics in model surgery
#69
levmckinney
closed
1 year ago
0
Code coverage reports are not working
#68
levmckinney
closed
1 year ago
0
added revision option to load; defaults to v0.0.4
#67
levmckinney
closed
1 year ago
0
Clean up the training code
#66
levmckinney
closed
1 year ago
2
Refactor CLI args Part 2 - Add type hinting to CLI args
#65
alexmlong
closed
1 year ago
0
Switch to using `simple-parsing`
#64
alexmlong
closed
1 year ago
4
A complete refactor of the plotting code
#63
levmckinney
closed
1 year ago
4
Remove spearmanr
#62
levmckinney
closed
1 year ago
0
Support TransformerLens
#61
UlisseMini
closed
1 year ago
1
Fix tokenization freeze
#60
levmckinney
closed
1 year ago
0
Refactor model surgery to no longer use heuristics
#59
levmckinney
closed
1 year ago
1
Get rid of torchrun; add --num-gpus argument to the main argparser
#58
norabelrose
closed
1 year ago
1
Something upstream appears to have broken the get_transformer_layers function.
#57
levmckinney
closed
1 year ago
0
get_final_layer_norm should throw if the LN can't be found
#56
norabelrose
closed
1 year ago
0
Refactors the `TunedLens` and `LogitLens` classes to now use the `Unembed` class
#55
levmckinney
closed
1 year ago
2
Fixes issue with running the interactive notebook on a cuda device
#54
levmckinney
closed
1 year ago
0
Problems running the notebook on GPU
#53
svenschultze
closed
1 year ago
1
Implement alternative decoder invert operation
#52
levmckinney
opened
1 year ago
1
Upgrade to pytorch 2.0
#51
levmckinney
closed
1 year ago
0
Fix the notebooks
#50
levmckinney
closed
1 year ago
0
What's 'white_box'?
#49
dbl001
closed
1 year ago
4
Allow for manual setting of tokenizer class
#48
norabelrose
closed
1 year ago
0
Remove vestigial CalibrationError stuff
#47
norabelrose
closed
1 year ago
0
CPU offload option
#46
norabelrose
closed
1 year ago
0
Use --revision for AutoTokenizer too
#45
norabelrose
closed
1 year ago
0
Dead code
#44
levmckinney
closed
1 year ago
0
Basic documentation using Sphinx
#43
levmckinney
closed
1 year ago
0
Lenses should use the Decoder class
#42
levmckinney
closed
1 year ago
1
Refactor Argument Parsers
#41
levmckinney
closed
1 year ago
0
Support for code coverage
#40
levmckinney
closed
1 year ago
3
Can not pull the docker image, is it public?
#39
xszheng2020
closed
1 year ago
2
Fixing acknowledgement links
#38
noodlefrenzy
closed
1 year ago
1
Better errors when artifact loading fails
#37
levmckinney
closed
1 year ago
0
The interactive notebook now supports colab
#36
levmckinney
closed
1 year ago
0
Previous
Next