INDEX

Explanations

keywords related to photography and videos

oai_token-act-pair · gpt-3.5-turbo

references to multimedia content, specifically photos and videos

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 extracted

-0.62

kan

-0.61

 default

-0.60

 meaning

-0.59

 margin

-0.59

 forest

-0.59

olin

-0.58

 capital

-0.57

 ability

-0.57

 increment

-0.56

POSITIVE LOGITS

PHOTOS

4.67

VIDEO

1.98

SPONSORED

1.51

MORE

1.49

RELATED

1.44

WATCH

1.30

SIGN

1.27

photos

1.26

Scroll

1.21

CLOSE

1.19

Activations Density 0.015%

keywords related to photography and videos

references to multimedia content, specifically photos and videos

No Comments

No Known Activations