INDEX

Explanations

numerical values that may be associated with specific data or statistics

oai_token-act-pair · gpt-3.5-turbo

special characters or graphical symbols

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

mania

-0.70

suff

-0.67

SON

-0.66

 Pitch

-0.64

metics

-0.64

 Kong

-0.64

lust

-0.63

 Shoot

-0.62

 Figures

-0.62

itu

-0.62

POSITIVE LOGITS

ãĥ¬

1.08

ãĤ·

0.94

pez

0.70

blem

0.66

 Attribution

0.64

onym

0.63

onyms

0.61

 hemisphere

0.60

istar

0.59

aspers

0.58

Activations Density 0.001%

numerical values that may be associated with specific data or statistics

special characters or graphical symbols

No Comments

No Known Activations