INDEX

Explanations

phrases related to computer and technology terms

oai_token-act-pair · gpt-3.5-turbo

instances of the letter 's'

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

dress

-0.68

orget

-0.64

docker

-0.63

mass

-0.61

EStream

-0.59

leans

-0.58

rolling

-0.57

 Rothschild

-0.57

 saline

-0.57

cake

-0.56

POSITIVE LOGITS

ELF

0.95

pecially

0.93

ullivan

0.89

 newest

0.87

selves

0.82

ources

0.80

own

0.78

udden

0.73

 flagship

0.73

wered

0.72

Activations Density 0.239%

phrases related to computer and technology terms

instances of the letter 's'

No Comments

No Known Activations