INDEX

Explanations

proper nouns related to individuals

oai_token-act-pair · gpt-3.5-turbo

repeated sequences of letters, particularly those that represent common name patterns or linguistic structures

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

catentry

-0.75

 lacking

-0.70

è¦ļéĨĴ

-0.68

DragonMagazine

-0.66

 Berm

-0.60

ALLY

-0.59

 Kinnikuman

-0.59

izoph

-0.58

 labyrinth

-0.57

 subt

-0.57

POSITIVE LOGITS

ghai

0.89

tty

0.86

inki

0.85

nit

0.84

puff

0.81

idan

0.81

bole

0.79

eki

0.78

oka

0.77

idi

0.77

Activations Density 0.060%

proper nouns related to individuals

repeated sequences of letters, particularly those that represent common name patterns or linguistic structures

No Comments

No Known Activations