INDEX

Explanations

people's names or proper nouns that contain a specific set of characters

oai_token-act-pair · gpt-3.5-turbo

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

llah

-0.69

OPLE

-0.65

Pwr

-0.60

merce

-0.59

 dism

-0.58

CLASSIFIED

-0.58

 thous

-0.57

 tumble

-0.57

 conscience

-0.57

 harrowing

-0.56

POSITIVE LOGITS

emort

1.03

uania

0.96

heed

0.95

ansk

0.88

gren

0.86

cious

0.86

emouth

0.85

ttle

0.82

uli

0.81

oyd

0.80

Activations Density 0.095%

people's names or proper nouns that contain a specific set of characters

No Comments

No Known Activations