INDEX

Explanations

the word "bored" or related variations

oai_token-act-pair · gpt-3.5-turbo

terms related to boredom and mundane situations

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

esville

-0.71

Ĥİ

-0.69

ortium

-0.65

cci

-0.62

bull

-0.60

ciation

-0.59

bow

-0.59

xon

-0.58

 Chun

-0.58

wark

-0.57

POSITIVE LOGITS

hower

0.91

ored

0.83

dit

0.79

66666666

0.78

andom

0.78

tsky

0.78

skirts

0.78

rite

0.76

ãĥ¯ãĥ³

0.75

hole

0.72

Activations Density 0.014%

the word "bored" or related variations

terms related to boredom and mundane situations

No Comments

No Known Activations