INDEX

Explanations

words and phrases that express concern or importance regarding various topics

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

emphasis on modifiers

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

protoimpl

-0.59

Autoritní

-0.57

multirow

-0.56

 rowCount

-0.54

})));

-0.54

存于互联网档案馆

-0.52

 myſelf

-0.51

 <>",

-0.50

 lenker

-0.50

Portale

-0.50

POSITIVE LOGITS

 particularly

1.03

Especially

0.94

 Particularly

0.94

 especially

0.93

 Especially

0.90

particularly

0.87

 especialmente

0.86

 particularmente

0.84

especially

0.79

 особенно

0.79

Activations Density 0.036%

words and phrases that express concern or importance regarding various topics

emphasis on modifiers

No Comments

No Known Activations