INDEX

Explanations

mentions of specific names, particularly those related to politics and medicine

oai_token-act-pair · gpt-3.5-turbo

proper nouns related to individuals and entities, particularly in a political or economic context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

kiss

-0.87

 Rays

-0.66

born

-0.66

 HERO

-0.63

 Austral

-0.63

 Quote

-0.62

friend

-0.62

 Lange

-0.60

 bearer

-0.60

burst

-0.60

POSITIVE LOGITS

owell

3.84

orr

1.64

umption

1.55

atel

1.39

uming

1.15

ennett

1.09

atton

1.03

agher

1.00

ermott

0.99

umers

0.95

Activations Density 0.030%

mentions of specific names, particularly those related to politics and medicine

proper nouns related to individuals and entities, particularly in a political or economic context

No Comments

No Known Activations