Adamliu1 SNLP_GCW issues

Adamliu1 / SNLP_GCW

3 stars 0 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Added harfmulness eval results.

#56 TheRootOf3 closed 8 months ago
0
Added flag to re-generate harmfulness eval results. Fixed plotting.

#55 TheRootOf3 closed 8 months ago
0
Add sanity check whether TruthfulQA is installed

#54 Willmish closed 3 months ago
1
Add total seen samples counting. Add some docs.

#53 Willmish closed 8 months ago
0
Beavertails_eval: Add flag for generating model answer with #### Question preifx

#52 Willmish closed 5 months ago
1
Remove saving model at step 0 as no longer needed.

#51 Willmish closed 8 months ago
1
feat: allow save_every for batch unlearning.

#50 Davidyz closed 8 months ago
1
Added a flag for not using lr_scheduler

#49 Adamliu1 closed 8 months ago
1
Fixed sequential batch size bug.

#48 TheRootOf3 closed 8 months ago
0
Changed the eval_results.sh script to reflect new unlearning script.

#47 TheRootOf3 closed 8 months ago
0
Format Python code with psf/black push

#46 github-actions[bot] closed 8 months ago
0
Implementation of the harmfulness evaluation following beavertails

#45 TheRootOf3 closed 8 months ago
0
Save model under unlearning at step 0.

#44 TheRootOf3 closed 8 months ago
0
Add beavertails-evaluation model getting answer from.

#43 Willmish closed 8 months ago
0
Add PKU-harmful evaluation to lm-eval-harness

#42 Willmish closed 8 months ago
2
Format Python code with psf/black push

#41 github-actions[bot] closed 8 months ago
0
Format Python code with psf/black push

#40 github-actions[bot] closed 8 months ago
0
Format Python code with psf/black push

#39 github-actions[bot] closed 8 months ago
0
Format Python code with psf/black push

#38 github-actions[bot] closed 8 months ago
0
Format Python code with psf/black push

#37 github-actions[bot] closed 8 months ago
0
Format Python code with psf/black push

#36 github-actions[bot] closed 8 months ago
0
Added license and authorship details. Removed unused code directories.

#35 TheRootOf3 closed 8 months ago
1
Format Python code with psf/black push

#34 github-actions[bot] closed 8 months ago
0
Restore bytedance implementation.

#33 Davidyz closed 8 months ago
0
Format Python code with psf/black push

#32 github-actions[bot] closed 8 months ago
0
Format Python code with psf/black push

#31 github-actions[bot] closed 8 months ago
0
Format Python code with psf/black push

#30 github-actions[bot] closed 8 months ago
0
fix: #14

#29 Davidyz closed 8 months ago
0
Format Python code with psf/black push

#28 github-actions[bot] closed 8 months ago
1
Format Python code with psf/black push

#27 github-actions[bot] closed 8 months ago
1
Format Python code with psf/black push

#26 github-actions[bot] closed 8 months ago
1
Format Python code with psf/black push

#25 github-actions[bot] closed 8 months ago
1
Add instructions for using `eval_framework_tasks/eval_results.sh` and `eval_framework_tasks/run_experiment.sh`

#24 Willmish closed 3 months ago
1
Format Python code with psf/black push

#23 github-actions[bot] closed 8 months ago
1
Format Python code with psf/black push

#22 github-actions[bot] closed 8 months ago
1
Format Python code with psf/black push

#21 github-actions[bot] closed 8 months ago
0
Format Python code with psf/black push

#20 github-actions[bot] closed 8 months ago
1
Format Python code with psf/black push

#19 github-actions[bot] closed 8 months ago
1
Added black formatter vscode setting and a github action.

#18 TheRootOf3 closed 8 months ago
1
Add black formatter gh action

#17 TheRootOf3 closed 8 months ago
0
Make `device` a required parameter in functions, wherever it is used.

#16 TheRootOf3 closed 3 months ago
1
MathQA broken for sequential unleaerning (fixme!!!)

#15 Willmish opened 8 months ago
2
Bug: Incorrect number of steps in lr scheduler steps for ByteDance

#14 TheRootOf3 closed 8 months ago
4
Refactor and housekeeping

#13 Davidyz closed 8 months ago
0
Add `lm-eval-harness` evaluation and analysis scripts.

#12 TheRootOf3 closed 8 months ago
1
Willmish/rebased mink logging + MathQA dataloader + blac formatting

#11 Willmish closed 8 months ago
0
FIx: Min-k% prob calculating after model step (possibly incorrect iwth gradient accumulation)

#10 Willmish closed 3 months ago
1
Add tokenizer as param to run_training_batch

#9 Willmish closed 3 months ago
3
Change checkpointing directory name and remove duplicated final checkpoint.

#8 TheRootOf3 closed 8 months ago
1
Clean and refactor `unlearn_harm.py`

#7 TheRootOf3 closed 8 months ago
0

Previous Next