INDEX

Explanations

phrases describing actions or events happening from a specific location

oai_token-act-pair · gpt-3.5-turbo

phrases indicating the source or perspective of information

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

merce

-0.72

ratulations

-0.70

idas

-0.68

peat

-0.68

xual

-0.66

blems

-0.65

ivariate

-0.64

itivity

-0.64

gradation

-0.64

potion

-0.63

POSITIVE LOGITS

 afar

1.55

 inside

1.00

 behind

0.98

 abroad

0.95

 upstairs

0.91

 nowhere

0.88

 whence

0.88

behind

0.87

 atop

0.86

 backstage

0.85

Activations Density 0.172%

phrases describing actions or events happening from a specific location

phrases indicating the source or perspective of information

No Comments

No Known Activations