INDEX

Explanations

phrases related to expressing opinions or making statements about various situations

oai_token-act-pair · gpt-3.5-turbo

conjunctions and phrases indicating relationships or connections

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

DX

-0.70

shift

-0.69

raits

-0.67

urry

-0.67

DragonMagazine

-0.65

uta

-0.64

 Powered

-0.64

amar

-0.64

iddy

-0.64

street

-0.62

POSITIVE LOGITS

 secondly

1.21

 hence

1.15

 therefore

1.14

 consequently

1.13

 furthermore

1.01

 thus

0.96

 blames

0.89

 vowed

0.85

 insisted

0.84

 believes

0.84

Activations Density 0.272%

phrases related to expressing opinions or making statements about various situations

conjunctions and phrases indicating relationships or connections

No Comments

No Known Activations