INDEX

Explanations

names of specific individuals, potentially related to news or articles

oai_token-act-pair · gpt-3.5-turbo

references to a specific individual named Upton

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 Investigator

-0.85

eworthy

-0.72

gans

-0.69

taboola

-0.67

ŀ

-0.66

 Investigators

-0.66

fm

-0.65

dk

-0.65

Ĳ

-0.64

ences

-0.64

POSITIVE LOGITS

 Upton

1.49

 Sinclair

0.89

 mosqu

0.86

minster

0.83

galitarian

0.76

stadt

0.76

 loft

0.73

 unpop

0.72

oun

0.71

 mathemat

0.71

Activations Density 0.007%

names of specific individuals, potentially related to news or articles

references to a specific individual named Upton

No Comments

No Known Activations