INDEX

Explanations

mentions of advertising

oai_token-act-pair · gpt-3.5-turbo

references to advertising

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

hew

-0.73

zh

-0.70

riot

-0.69

ious

-0.67

 Sequence

-0.66

orable

-0.66

 Stevens

-0.66

rior

-0.66

ave

-0.64

aves

-0.64

POSITIVE LOGITS

 advertising

3.61

 Advertising

2.67

 advertisements

2.58

ads

2.36

 advertisement

2.31

 marketing

2.13

 advertisers

2.12

 advert

1.96

 commercials

1.96

 advertis

1.96

Activations Density 0.005%

mentions of advertising

references to advertising

No Comments

No Known Activations