kmeng01/rome

Issues

• GPT2-medium stats not accessible.

  #49 opened a month ago by ansh997
  0

• Code for computing the right vector for the rank-1 update

  #32 opened 2 years ago by QZH-777
  1

• TypeError when running causal_trace.ipynb locally

  #48 opened 3 months ago by LuoXiaoxi-cxq
  0

• Code for Average Plot over Multiple Runs

  #30 opened 2 years ago by aflah02
  2

• An interesting phenomenon

  #46 opened 5 months ago by XeeKee
  1

• How to apply this methods to a new LMs without fine-tune, like a bert?

  #28 opened 2 years ago by sev777
  2

• Tensorflow Support

  #47 opened 5 months ago by lxianl455
  0

• Does the number of samples not affect casual tracing?

  #45 opened 6 months ago by Jo-won
  1

• Possible bug in fine-tuning baseline implementation

  #41 opened a year ago by drd13
  1

• Failed on causal tracing

  #43 opened 10 months ago by JX-Zhang98
  1

• installation

  #42 opened 10 months ago by ZHANGJINKUI
  1

• Conda setup issue

  #36 opened a year ago by matteorr
  1

• A question about casual tracing code

  #40 opened a year ago by Zijian007
  0

• Why you choose layer 18 as a edits layer

  #37 opened a year ago by Robin-WZQ
  0

• What is the 'simple' principle of ROME?

  #34 opened a year ago by guotong1988
  0

• prob contributions by individual neuron rather than layer

  #7 opened 3 years ago by nathanneuro
  2

• What is the performance of ROME at editing GPT-NeoX-20b?

  #31 opened 2 years ago by Zce1112zslx
  0

• How do you train KE and MEND with CounterFact?

  #29 opened 2 years ago by Zce1112zslx
  0

• Documentation for hparams?

  #27 opened 2 years ago by moyix
  2

• CUDA out of memory error on an 11 GB GPU: Any easy ways to use multiple GPUs?

  #26 opened 2 years ago by yashjakhotiya
  0

• Model- and data-dependent hyperparameters

  #25 opened 2 years ago by negar-foroutan
  3

• How do you calculate Efficacy Score for multi-token words in your counterfact dataset?

  #24 opened 2 years ago by Zce1112zslx
  1

• Any suggestions for extending this work to edit values?

  #18 opened 2 years ago by QuintinPope
  1

• Generating weights for efk/mend for new model

  #12 opened 2 years ago by salemohamedo
  1

• Best way to specify a non-standard cache directory

  #13 opened 2 years ago by QuintinPope
  1

• ImportError: cannot import name 'Literal' from 'typing' (/usr/lib/python3.7/typing.py)

  #23 opened 2 years ago by JayThibs
  2

• CUDA Out of Memory Error on Colab

  #10 opened 3 years ago by aswathyajith
  3

• T5 support

  #2 opened 3 years ago by puraminy
  2

• Confusion about pre/post rewrite probabilities of target true and target new

  #6 opened 3 years ago by salemohamedo
  5