INDEX

Explanations

phrases that indicate events happening before or leading up to another event

oai_token-act-pair · gpt-3.5-turbo

the word "prior" and its variations, indicating a focus on preceding events or conditions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

aden

-0.68

rosso

-0.62

VILLE

-0.62

_-

-0.61

 Elves

-0.57

SET

-0.57

=-=-=-=-=-=-=-=-

-0.57

RO

-0.56

FOR

-0.56

Sort

-0.55

POSITIVE LOGITS

itiz

1.53

ities

1.20

itized

1.16

to

0.96

 thereto

0.96

ity

0.84

ebin

0.79

izations

0.79

ITY

0.78

itÃ©

0.78

Activations Density 0.019%

phrases that indicate events happening before or leading up to another event

the word "prior" and its variations, indicating a focus on preceding events or conditions

No Comments

No Known Activations