INDEX

Explanations

phrases related to duration or length of time

oai_token-act-pair · gpt-3.5-turbo

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

ĪĴ

-0.64

ymes

-0.63

ãĤ¼ãĤ¦ãĤ¹

-0.62

ulence

-0.59

ãĥ¯ãĥ³

-0.59

essen

-0.59

uterte

-0.58

leck

-0.56

ESA

-0.55

ãĤµ

-0.54

POSITIVE LOGITS

 until

1.59

 till

1.50

 indefinitely

1.46

 forever

1.41

until

1.35

 since

1.27

 awhile

1.21

 longest

1.10

 longer

1.09

 continuously

1.04

Activations Density 1.052%

phrases related to duration or length of time

No Comments

No Known Activations