INDEX

Explanations

symbols followed by a statement or conclusion

oai_token-act-pair · gpt-3.5-turbo

colons and statements introducing significant points or questions

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.hook_resid_pre

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

izons

-0.78

oud

-0.72

ema

-0.71

gered

-0.71

-0.69

agon

-0.69

onet

-0.69

ebin

-0.68

leased

-0.68

iban

-0.66

POSITIVE LOGITS

 namely

0.96

 Whenever

0.85

 Provided

0.83

 Anyone

0.83

 Forget

0.80

 Where

0.79

 When

0.79

 Firstly

0.79

 Someone

0.78

 Until

0.78

Activations Density 0.124%

symbols followed by a statement or conclusion

colons and statements introducing significant points or questions

No Comments

No Known Activations