INDEX

Explanations

visual content or imagery

oai_token-act-pair · gpt-3.5-turbo

instances of the word "images."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

cffff

-0.78

woods

-0.76

merce

-0.71

wood

-0.70

ggles

-0.70

dule

-0.69

sole

-0.68

secut

-0.68

upuncture

-0.68

odo

-0.67

POSITIVE LOGITS

 images

1.07

 galleries

1.03

ynt

1.01

ynthesis

1.00

 Images

0.97

 depicting

0.97

 IMAGES

0.96

 photographed

0.93

 imagery

0.92

Images

0.91

Activations Density 0.020%

visual content or imagery

instances of the word "images."

No Comments

No Known Activations