INDEX

Explanations

phrases related to reading additional content or instructions online

oai_token-act-pair · gpt-3.5-turbo

instances of clickable content or calls to action

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

uyomi

-0.71

Tur

-0.68

agement

-0.67

arbon

-0.65

attled

-0.64

Decre

-0.63

 surrog

-0.62

è¦ļéĨĴ

-0.62

 Acceler

-0.62

 Improvement

-0.62

POSITIVE LOGITS

 here

1.15

 HERE

1.00

here

1.00

 Here

0.93

 there

0.84

 herein

0.83

Here

0.81

 ours

0.79

chens

0.78

 abroad

0.77

Activations Density 0.209%

phrases related to reading additional content or instructions online

instances of clickable content or calls to action

No Comments

No Known Activations