INDEX

Explanations

images described within the text

oai_token-act-pair · gpt-3.5-turbo

repeated references to images

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

hips

-0.72

asonable

-0.66

termination

-0.65

ospons

-0.65

Ear

-0.64

 Summers

-0.62

behind

-0.62

verning

-0.62

ieves

-0.61

burg

-0.60

POSITIVE LOGITS

 image

3.68

image

2.57

 Image

2.48

 images

2.47

 photograph

1.93

 picture

1.89

Image

1.89

 photo

1.82

 imagery

1.81

images

1.79

Activations Density 0.018%

images described within the text

repeated references to images

No Comments

No Known Activations

images described within the text

repeated references to images

No Comments

No Known Activations