INDEX

Explanations

sequences of underscores or whitespace characters

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

programming terms and code

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

New Auto-Interp

Top Features by Cosine Similarity

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

Configuration

google/gemma-scope-9b-it-res/layer_9/width_131k/average_l0_121

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.9.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

SourceChecksum

-0.79

 Administrativna

-0.66

CloseOperation

-0.64

 Italijanski

-0.64

ьажоргаш

-0.64

potranspiration

-0.63

 مرئيه

-0.63

ThemeOverlay

-0.61

antMatchers

-0.61

cerpt

-0.60

POSITIVE LOGITS

 NSCoder

0.47

didSet

0.34

 éter

0.31

 изменения

0.31

 عض

0.30

 change

0.29

 habitudes

0.29

Succeed

0.28

 insufic

0.28

Hostname

0.28

Activations Density 0.002%

sequences of underscores or whitespace characters

programming terms and code

No Comments

No Known Activations