INDEX

Explanations

affirmative modifiers expressing certainty and positivity

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

emphasizing positive qualities

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

 &___

-0.40

ScopeManager

-0.39

 вдруг

-0.36

 trânsito

-0.34

 secrétaire

-0.33

 drame

-0.33

Попис

-0.33

dafx

-0.31

 aislado

-0.31

onomía

-0.31

POSITIVE LOGITS

 worthy

0.71

 truly

0.66

 worth

0.62

Javadoc

0.62

 deserving

0.60

 true

0.56

 deserve

0.56

 layak

0.55

mpion

0.55

zzá

0.54

Activations Density 0.007%

affirmative modifiers expressing certainty and positivity

emphasizing positive qualities

No Comments

No Known Activations