INDEX

Explanations

punctuation and numerical values in data

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

symbols and numbers in parentheses

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

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

Спасылкі

-0.70

OGND

-0.68

 lenker

-0.67

########.

-0.66

 Administrativna

-0.66

 Wikimedijinoj

-0.65

 gyhoeddwyd

-0.65

 kasarigan

-0.63

errHandler

-0.62

 ModelExpression

-0.60

POSITIVE LOGITS

 soñar

0.39

 actuels

0.37

 vuestro

0.36

 présentés

0.36

 folha

0.36

 hechos

0.36

 terbukti

0.35

 indivíduos

0.35

 faibles

0.35

 fraî

0.35

Activations Density 0.309%

punctuation and numerical values in data

symbols and numbers in parentheses

No Comments

No Known Activations