INDEX

Explanations

comparative terms that indicate differences in quality or quantity

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

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

DataAnnotations

-0.50

Cyfarwyddwr

-0.47

sidemargin

-0.47

仅

-0.43

libft

-0.42

amos

-0.42

 irra

-0.42

ragt

-0.42

 constantly

-0.42

gnings

-0.41

POSITIVE LOGITS

 than

1.22

 happier

0.85

 niż

0.82

 THAN

0.80

 harder

0.80

 brighter

0.80

 smoother

0.79

 quicker

0.78

 darker

0.78

 funnier

0.78

Activations Density 0.588%

comparative terms that indicate differences in quality or quantity

No Comments

No Known Activations