INDEX

Explanations

phrases indicating high frequency or prevalence of concepts or phenomena

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

New Auto-Interp

Top Features by Cosine Similarity

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

Configuration

google/gemma-scope-2b-pt-res/layer_9/width_16k/average_l0_73

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.9.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

ParallelGroup

-0.56

עם

-0.53

 THEIR

-0.50

 InputDecoration

-0.49

weren

-0.49

Separate

-0.48

 their

-0.47

roff

-0.47

noscript

-0.46

 Herren

-0.46

POSITIVE LOGITS

 commonly

1.07

 Commonly

1.01

commonly

0.97

 common

0.88

常見

0.87

 widespread

0.86

 popular

0.86

 häufig

0.86

popular

0.86

 prevalent

0.86

Activations Density 0.404%

phrases indicating high frequency or prevalence of concepts or phenomena

No Comments

No Known Activations