INDEX

Explanations

phrases related to technical terms and programming

oai_token-act-pair · gpt-3.5-turbo

specific single-letter or short abbreviations and associated terms

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

 Dennis

-0.95

 DRAG

-0.88

 Woodward

-0.86

Daw

-0.83

Dol

-0.83

Mae

-0.79

DAR

-0.76

 Wesley

-0.73

è¯

-0.73

 Cary

-0.73

POSITIVE LOGITS

IT

1.49

ite

1.48

ites

1.44

it

1.40

iter

1.30

ITE

1.26

ito

1.24

ita

1.19

its

1.17

iting

1.16

Activations Density 0.184%

phrases related to technical terms and programming

specific single-letter or short abbreviations and associated terms

No Known Activations