INDEX

Explanations

phrases or names related to individuals in various contexts

oai_token-act-pair · gpt-3.5-turbo

names of individuals, particularly those of people involved in commentary or analysis

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

eenth

-0.88

uyomi

-0.74

ivid

-0.68

sburgh

-0.67

 tomat

-0.66

 Jagu

-0.64

angular

-0.64

arov

-0.63

llular

-0.63

holders

-0.62

POSITIVE LOGITS

zen

1.05

kel

0.96

thal

0.92

ck

0.85

 Dres

0.84

culated

0.84

con

0.83

zer

0.81

Ã§

0.81

tha

0.81

Activations Density 0.022%

phrases or names related to individuals in various contexts

names of individuals, particularly those of people involved in commentary or analysis

No Comments

No Known Activations