INDEX

Explanations

references to numerical values and quantifiable data in various contexts

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

-0.46

anstalt

-0.44

QObject

-0.41

 जाता

-0.41

 оригіналу

-0.41

//

-0.40

@[+][

-0.39

UniformLocation

-0.39

 worse

-0.39

PullParser

-0.38

POSITIVE LOGITS

Remaining

0.49

 remaining

0.49

 Мексичка

0.48

satunya

0.47

 którzy

0.47

 beiden

0.47

 linkovi

0.47

 pozosta

0.46

 ktorí

0.45

 endregion

0.45

Activations Density 0.181%

references to numerical values and quantifiable data in various contexts

No Comments

No Known Activations