INDEX

Explanations

words related to power and domination

oai_token-act-pair · gpt-3.5-turbo

occurrences of the letter 'p'

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 GOODMAN

-0.94

ãĤ´ãĥ³

-0.89

æ©Ł

-0.85

DragonMagazine

-0.81

FTWARE

-0.80

MpServer

-0.78

ãĤ¼ãĤ¦ãĤ¹

-0.78

 compe

-0.76

 Halls

-0.74

æ©

-0.74

POSITIVE LOGITS

uffed

1.22

aired

1.12

orters

1.12

airs

1.11

appy

1.11

ussy

1.10

ascal

1.10

orter

1.09

udd

1.08

atern

1.08

Activations Density 0.032%

words related to power and domination

occurrences of the letter 'p'

No Comments

No Known Activations