INDEX

Explanations

expressions of gratitude

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

expression of thanks

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

pincode

-0.45

`.`

-0.44

or

-0.44

 Coimbatore

-0.43

 individual

-0.42

 Museum

-0.42

 Woman

-0.41

ellón

-0.41

 essay

-0.41

 canst

-0.41

POSITIVE LOGITS

Thanks

0.95

 Thanks

0.93

thanks

0.92

THANKS

0.90

 THANKS

0.86

 thanks

0.79

Thx

0.72

thx

0.72

Thx

0.70

gracias

0.67

Activations Density 0.008%

expressions of gratitude

expression of thanks

No Comments

No Known Activations

expressions of gratitude

expression of thanks

No Comments

No Known Activations