INDEX

Explanations

the word "any" followed by a number that represents various technical functions or features

oai_token-act-pair · gpt-3.5-turbo

the word "any" used in various contexts

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

endas

-0.78

illard

-0.70

uble

-0.70

combe

-0.69

pees

-0.68

gers

-0.68

forth

-0.66

vier

-0.65

bows

-0.65

flation

-0.64

POSITIVE LOGITS

THING

1.26

WHERE

1.13

where

1.08

 conceivable

1.05

 kind

1.00

body

0.99

 sort

0.95

ONE

0.95

 imaginable

0.92

 other

0.91

Activations Density 0.071%

the word "any" followed by a number that represents various technical functions or features

the word "any" used in various contexts

No Comments

No Known Activations