INDEX

Explanations

prominent names of politicians and public figures in discussions or mentions

oai_token-act-pair · gpt-3.5-turbo

proper nouns, specifically names of individuals or political figures

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

berra

-0.69

olulu

-0.63

cgi

-0.63

ARS

-0.62

Mount

-0.62

lihood

-0.62

 Flavoring

-0.62

ðĿ

-0.60

oxide

-0.60

POSITIVE LOGITS

 detractors

0.89

 himself

0.88

 supporters

0.84

 aides

0.81

's

0.77

 campaigned

0.76

 backers

0.76

 tweeted

0.75

 supporter

0.74

 advisor

0.74

Activations Density 0.290%

prominent names of politicians and public figures in discussions or mentions

proper nouns, specifically names of individuals or political figures

No Comments

No Known Activations