INDEX

Explanations

proper nouns, specifically related to political figures and individuals

oai_token-act-pair · gpt-3.5-turbo

mentions of specific historical figures and their related contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ardless

-0.89

Oracle

-0.79

toe

-0.77

 Territories

-0.77

ebook

-0.73

 Printed

-0.71

 Trooper

-0.71

Spread

-0.70

Tube

-0.69

 Olsen

-0.68

POSITIVE LOGITS

escription

0.97

VIDIA

0.88

olulu

0.88

ilater

0.85

interstitial

0.84

 belie

0.84

yssey

0.83

iatus

0.80

 immersion

0.77

anan

0.77

Activations Density 0.016%

proper nouns, specifically related to political figures and individuals

mentions of specific historical figures and their related contexts

No Comments

No Known Activations