INDEX

Explanations

the word "the" in sequences of words

oai_token-act-pair · gpt-3.5-turbo

frequent instances of the word "the"

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

thood

-0.81

 suppose

-0.69

ÄŁ

-0.69

replace

-0.68

iatus

-0.66

ItemThumbnailImage

-0.65

FontSize

-0.64

Enjoy

-0.63

elaide

-0.62

amaz

-0.62

POSITIVE LOGITS

 hardest

1.25

 entire

1.23

 same

1.20

 entirety

1.10

ses

1.09

 slightest

1.08

 whole

1.06

 remainder

1.03

 toughest

1.02

 proverbial

0.95

Activations Density 0.244%

the word "the" in sequences of words

frequent instances of the word "the"

No Comments

No Known Activations