INDEX

Explanations

names of people and potentially organizations

oai_token-act-pair · gpt-3.5-turbo

proper nouns and names

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 Carnage

-0.67

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

-0.66

 lesbians

-0.65

 Masquerade

-0.63

 Somali

-0.63

 Sigma

-0.62

[&

-0.62

cknowled

-0.62

 Skydragon

-0.62

initialized

-0.62

POSITIVE LOGITS

odore

0.75

dinand

0.75

romeda

0.69

alin

0.69

ison

0.68

entin

0.67

berman

0.67

rick

0.67

undy

0.65

espie

0.65

Activations Density 0.217%

names of people and potentially organizations

proper nouns and names

No Comments

No Known Activations