INDEX

Explanations

phrases related to publicity stunts

oai_token-act-pair · gpt-3.5-turbo

various uses of the word "stunt" in different contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

vironment

-0.70

ishops

-0.66

oice

-0.65

icrobial

-0.64

oan

-0.64

 Chapters

-0.64

uni

-0.63

soType

-0.62

inda

-0.61

 Sett

-0.61

POSITIVE LOGITS

 stunt

0.93

 stunts

0.92

hered

0.90

strip

0.78

olicy

0.76

monkey

0.75

nel

0.74

crew

0.73

 sled

0.73

enegger

0.71

Activations Density 0.020%

phrases related to publicity stunts

various uses of the word "stunt" in different contexts

No Comments

No Known Activations