INDEX

Explanations

mentions of things being or becoming popular

oai_token-act-pair · gpt-3.5-turbo

references to the concept of popularity

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

Â¯Â¯

-0.86

erm

-0.82

ibur

-0.79

abol

-0.76

alk

-0.74

cise

-0.74

holes

-0.71

htaking

-0.69

gans

-0.69

rib

-0.68

POSITIVE LOGITS

 popularity

1.11

popular

0.91

 Popular

0.86

yip

0.84

 ratings

0.80

 ubiqu

0.77

 unpopular

0.75

 renown

0.75

é¾įå¥ĳå£«

0.75

 diffusion

0.74

Activations Density 0.009%

mentions of things being or becoming popular

references to the concept of popularity

No Comments

No Known Activations