INDEX

Explanations

proper nouns related to politics, economics, and historical events

oai_token-act-pair · gpt-3.5-turbo

references to historical political figures and events, particularly those related to conservatism and the Titanic

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

externalActionCode

-0.78

¥µ

-0.72

Hop

-0.64

 Aether

-0.64

Asia

-0.64

Interested

-0.63

Tek

-0.63

sure

-0.62

CAST

-0.62

reads

-0.61

POSITIVE LOGITS

 Thatcher

1.26

ite

0.84

ites

0.83

ism

0.82

chev

0.80

nikov

0.80

imar

0.76

 nostalg

0.74

isers

0.74

isms

0.74

Activations Density 0.010%

proper nouns related to politics, economics, and historical events

references to historical political figures and events, particularly those related to conservatism and the Titanic

No Comments

No Known Activations