INDEX

Explanations

phrases related to analyzing and summarizing information

oai_token-act-pair · gpt-3.5-turbo

phrases related to analysis and review

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

ief

-0.73

rib

-0.72

icum

-0.71

IUM

-0.69

Pg

-0.69

riber

-0.69

END

-0.68

achable

-0.68

ONG

-0.67

iverse

-0.65

POSITIVE LOGITS

why

1.33

how

1.33

 some

1.07

 what

1.03

why

0.96

 aspects

0.95

 ways

0.94

 exactly

0.94

 strategies

0.91

WHY

0.89

Activations Density 0.196%

phrases related to analyzing and summarizing information

phrases related to analysis and review

No Comments

No Known Activations