INDEX

Explanations

images or descriptions explicitly mentioning individuals or objects in a visual context

oai_token-act-pair · gpt-3.5-turbo

instances of the word "pictured" and its variations, indicating visual references within the text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

cession

-0.88

merce

-0.83

angan

-0.76

idity

-0.74

vernment

-0.74

ced

-0.72

alities

-0.70

die

-0.70

stood

-0.68

ggles

-0.67

POSITIVE LOGITS

 pictured

1.02

 Thumbnails

0.97

 IMAGES

0.84

above

0.83

 Pict

0.81

amera

0.79

Pict

0.78

 below

0.77

 photographed

0.76

 Gallery

0.75

Activations Density 0.023%

images or descriptions explicitly mentioning individuals or objects in a visual context

instances of the word "pictured" and its variations, indicating visual references within the text

No Comments

No Known Activations