INDEX

Explanations

references to written communication and formal expressions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 38-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_38/width_16k/average_l0_128

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.38.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

 <=",

-0.79

PostInfinity

-0.71

 AssemblyCulture

-0.71

 müſſen

-0.70

 Numerade

-0.69

 Ferner

-0.68

/**

-0.67

 laſſen

-0.67

 ویکی‌پدی

-0.65

 zuſammen

-0.65

POSITIVE LOGITS

mal

0.50

 mega

0.43

nix

0.43

^^

0.36

 halt

0.36

hab

0.33

 😅

0.33

nix

0.32

dir

0.32

rum

0.32

Activations Density 0.073%

references to written communication and formal expressions

No Comments

No Known Activations