INDEX

Explanations

moral and ethical dilemmas, possibly related to religion

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 targeted

-0.61

 telecommunications

-0.60

 marching

-0.56

 machine

-0.53

 aiming

-0.53

 bulldo

-0.52

 basket

-0.52

 broadcasting

-0.52

 threatened

-0.52

 broadband

-0.51

POSITIVE LOGITS

 Conclusion

0.93

Conclusion

0.83

laughs

0.80

 Lastly

0.80

APTER

0.75

________________

0.74

 CONTIN

0.71

CHAPTER

0.71

________________________________

0.71

Lastly

0.71

Activations Density 0.409%

moral and ethical dilemmas, possibly related to religion

No Comments

No Known Activations