INDEX

Explanations

terms related to specific proper nouns, likely names of individuals and businesses

oai_token-act-pair · gpt-3.5-turbo

references to names and terms associated with specific individuals or entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

wash

-0.75

osph

-0.72

tion

-0.71

 Chomsky

-0.68

mer

-0.67

mers

-0.66

pot

-0.65

brain

-0.63

tor

-0.63

hler

-0.62

POSITIVE LOGITS

eties

0.81

DragonMagazine

0.76

 footh

0.76

esse

0.75

geoning

0.74

ELY

0.72

 Dull

0.71

hest

0.70

ashtra

0.70

naires

0.70

Activations Density 0.066%

terms related to specific proper nouns, likely names of individuals and businesses

references to names and terms associated with specific individuals or entities

No Comments

No Known Activations