INDEX

Explanations

captions or descriptions marked for enlargement in images

oai_token-act-pair · gpt-3.5-turbo

commands related to enlarging 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

 Rowe

-0.67

 cunt

-0.65

 cant

-0.63

 submission

-0.63

 good

-0.62

 decent

-0.62

 defence

-0.61

wil

-0.61

 monster

-0.59

 wicked

-0.59

POSITIVE LOGITS

Enlarge

5.40

toggle

2.94

NPR

1.64

SPONSORED

1.42

Newsletter

1.42

externalActionCode

1.32

DragonMagazine

1.17

Reviewer

1.13

 WATCHED

1.11

Interstitial

1.10

Activations Density 0.019%

captions or descriptions marked for enlargement in images

commands related to enlarging images

No Comments

No Known Activations