INDEX

Explanations

instances of the word 'real' signifying authenticity or truthfulness

oai_token-act-pair · gpt-3.5-turbo

instances of the word "real" indicating authenticity or truthfulness

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ebook

-0.82

berries

-0.76

xual

-0.75

asers

-0.66

MENTS

-0.64

 Travels

-0.64

heid

-0.64

mph

-0.64

ACK

-0.63

utions

-0.63

POSITIVE LOGITS

ignment

1.26

 estate

1.15

polit

1.12

igning

1.11

igned

1.03

isation

0.99

estate

0.97

DonaldTrump

0.95

istically

0.94

isable

0.88

Activations Density 0.027%

instances of the word 'real' signifying authenticity or truthfulness

instances of the word "real" indicating authenticity or truthfulness

No Comments

No Known Activations