INDEX

Explanations

phrases related to official documents or statements

oai_token-act-pair · gpt-3.5-turbo

references to statements and their sources

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

ciples

-0.75

hene

-0.73

lied

-0.72

gradation

-0.71

merce

-0.69

ãĤ¦ãĤ¹

-0.69

iencies

-0.69

mun

-0.68

cipl

-0.67

important

-0.67

POSITIVE LOGITS

 afar

0.95

 Washington

0.83

 inside

0.80

 whence

0.78

 abroad

0.75

 Denmark

0.73

 Mayo

0.73

 Lans

0.73

 Seattle

0.73

 Minneapolis

0.72

Activations Density 0.123%

phrases related to official documents or statements

references to statements and their sources

No Comments

No Known Activations