INDEX

Explanations

capitalized words conveying strong emotions or significance

oai_token-act-pair · gpt-3.5-turbo

words and phrases that indicate emotional or evaluative responses

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

Ag

-0.66

weed

-0.66

-0.62

 âĢº

-0.62

Engineers

-0.60

 Tibetan

-0.54

Company

-0.53

Matrix

-0.53

bert

-0.53

 Â®

-0.53

POSITIVE LOGITS

rul

0.70

KE

0.69

hower

0.65

)=(

0.65

ODY

0.64

ebted

0.62

 CLIENT

0.62

 condem

0.60

eatures

0.59

ilogy

0.59

Activations Density 2.067%

capitalized words conveying strong emotions or significance

words and phrases that indicate emotional or evaluative responses

No Comments

No Known Activations