INDEX

Explanations

phrases related to media attention and promotion

oai_token-act-pair · gpt-3.5-turbo

references to publicity in various 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

lements

-0.87

tein

-0.81

inite

-0.69

azes

-0.69

xit

-0.68

atchewan

-0.67

artment

-0.67

llan

-0.64

ling

-0.64

onian

-0.63

POSITIVE LOGITS

 publicity

0.95

 tremend

0.90

 stunts

0.87

quickShipAvailable

0.84

 endorsements

0.80

campaign

0.78

 stunt

0.76

æ©

0.74

orial

0.73

GROUP

0.70

Activations Density 0.036%

phrases related to media attention and promotion

references to publicity in various contexts

No Comments

No Known Activations