INDEX

Explanations

references to 'all' and collective terms

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

the word "all" or quantitative expressions about entirety.

oai_token-act-pair · claude-3-5-haiku-20241022 Triggered by @neilrathi

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_16k/average_l0_68

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Jacobsen

-0.47

 device

-0.45

 scheme

-0.45

ed

-0.43

 programme

-0.43

 Jacobson

-0.42

ised

-0.41

<?=

-0.41

hamshire

-0.41

 febru

-0.41

POSITIVE LOGITS

All

1.22

All

1.20

all

1.17

setAll

1.11

ALL

1.11

 Semua

1.02

all

0.99

ALL

0.98

 semua

0.93

 GetAll

0.89

Activations Density 0.155%

references to 'all' and collective terms

the word "all" or quantitative expressions about entirety.

No Comments

No Known Activations