INDEX

Explanations

options being considered or evaluated in diverse contexts

oai_token-act-pair · gpt-3.5-turbo

phrases discussing choices or alternatives

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

zi

-0.78

aph

-0.78

hemat

-0.76

abol

-0.73

wind

-0.68

organ

-0.68

ews

-0.68

bug

-0.67

storm

-0.67

ric

-0.67

POSITIVE LOGITS

 options

1.55

 choices

1.15

 alternatives

1.09

options

1.08

 option

1.05

 Options

1.00

 avenues

0.95

DragonMagazine

0.92

ensical

0.91

atives

0.86

Activations Density 0.011%

options being considered or evaluated in diverse contexts

phrases discussing choices or alternatives

No Comments

No Known Activations