likenneth honest_llama issues

likenneth / honest_llama

Inference-Time Intervention: Eliciting Truthful Answers from a Language Model

MIT License

478 stars 37 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Loading checkpoint of saved intervened model takes long time

#45 jeffreyzhanghc opened 1 week ago
4
Clean iti replicate

#44 likenneth closed 2 months ago
0
Cannot reproduce the probe result (Figure 2: (A) in the paper)

#43 jiahaozhenbang closed 1 month ago
4
Clarification on where to intervene

#42 CheongWoong closed 2 months ago
4
Upload ITI baked-in models to HuggingFace

#41 jujipotle closed 3 months ago
0
Will you publish the Llama3 fine-tuned model?

#40 aryopg closed 3 months ago
1
Will you publish the Llama3 fine-tuned model?

#39 aryopg closed 3 months ago
4
Cleaned up replication of ITI on Llama3

#38 jujipotle closed 4 months ago
0
ask about the insight behind ασθ

#37 NieSYsc20 closed 4 months ago
1
Replicate iti llama3

#36 jujipotle closed 4 months ago
0
Issues related to reproducing the results of the paper

#35 wytbwytb closed 5 months ago
3
Clarifications on table 5 of paper

#34 itsmemala closed 5 months ago
4
Code for CCS

#33 itsmemala closed 5 months ago
1
Query regarding dimensions of activations

#32 itsmemala closed 9 months ago
1
Interesting work! Providing an additional convenient way of reproducing the results!

#31 frankaging closed 10 months ago
1
Discrepancy in Reproducing Results with llama-7B on TriviaQA Dataset

#30 XianfengJiao closed 5 months ago
5
llama2_chat_13B and hyperparameters sweep

#29 marov closed 11 months ago
0
How to calculate the gap between generation accuracy and probe accuracy, which is 40% mentioned in the paper?

#28 DLiquor closed 11 months ago
2
issue in validate_2fold.py ordering csv by huggingface order

#27 tianlwang closed 11 months ago
7
Why does memory accumulate and ultimately cause overflow when running get_activations.py?

#26 Renpf2022 closed 7 months ago
4
Does ITI support Qwen？

#25 menghonghan closed 11 months ago
1
Llama2 chat 70b support

#24 marov closed 12 months ago
13
Inquiry on the GPT-judge cost and potential subsititutions

#23 night-chen closed 7 months ago
3
False Answer in OamPatel/iti_trivia_qa_val

#22 Vicent0205 closed 1 year ago
3
Support llama2 series models？

#21 skykiseki closed 1 year ago
2
Cannot replicate results on judge and info metric

#20 fabrahman closed 7 months ago
6
why u chose ‘tqa_gen_end_q’ to compute std and not ‘tqa_mc2’ to compute std.

#19 thyywr759 closed 1 year ago
1
Question: doess INI support other kinds model like mpt or baichuan?

#18 Mewral closed 1 year ago
1
Query on result of Vicuna and Alpaca

#17 jongjyh closed 1 year ago
2
Update README.md

#16 notrichardren closed 1 year ago
0
Why do we use `llama.*` instead of HuggingFace's llama?

#15 RylanSchaeffer closed 1 year ago
1
Comparison with other tuning methods

#14 FLLLIGHT closed 1 year ago
1
Cannot find truthful_qa.py

#13 LebronX closed 1 year ago
3
Saving the model after shifting activations

#12 A-Raafat closed 1 year ago
1
Difference between tqag_gen_end_q and tqa_gen?

#11 jongjyh closed 1 year ago
1
The result of the code doesn't match the result in the paper

#10 CaoYiqingT closed 1 year ago
2
How to use tqa_gen and tqa_end_end_q?

#9 CaoYiqingT closed 1 year ago
2
Which part of the paper does tqa_gen_end_q correspond to？

#8 CaoYiqingT closed 1 year ago
2
what is self_attn.head_out

#7 JianqiaoLu closed 1 year ago
3
Potential Data Leakage in Probes Training

#6 jongjyh closed 1 year ago
7
disagreement about truthful qa results

#5 Vicent0205 closed 1 year ago
3
inquery of visualizing result on the paper

#4 jongjyh closed 1 year ago
3
inquery on equaltion of paper

#3 jongjyh closed 1 year ago
1
validation code seems to have only few-shot settings

#2 Maxlinn closed 1 year ago
3
Ask for providing the GPT-4 generated false answers for NQ/TriviaQA

#1 voidism closed 1 year ago
6