INDEX

Explanations

terms related to 'high level' concepts or hierarchies

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

high-level and large-scale

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

RenderAtEndOf

-0.49

 Ajoutez

-0.44

vœ

-0.43

 Connectez

-0.42

 doença

-0.41

 mijne

-0.41

 Selecciona

-0.40

 botas

-0.40

 hambre

-0.40

 Garantía

-0.39

POSITIVE LOGITS

nonatomic

0.62

 NSObject

0.56

 للاسماء

0.55

 Advanced

0.54

 informal

0.54

 advanced

0.53

isticated

0.53

 Highly

0.52

irited

0.52

 highly

0.51

Activations Density 0.012%

terms related to 'high level' concepts or hierarchies

high-level and large-scale

No Comments

No Known Activations