INDEX

Explanations

programming language keywords and structure

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

pub fn client

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

AnimationsModule

-0.35

 strony

-0.35

NINGS

-0.34

empêcher

-0.33

 shortage

-0.33

 custa

-0.33

fD

-0.33

Students

-0.33

 Bestimmungen

-0.32

//

-0.32

POSITIVE LOGITS

 nahilalakip

0.80

 AssemblyCulture

0.77

 noDo

0.69

 незавершена

0.68

 kasarigan

0.66

uxxxx

0.65

Билгалдахарш

0.64

 تضيفلها

0.64

Tikang

0.63

лтемелер

0.61

Activations Density 0.005%

programming language keywords and structure

pub fn client

No Comments

No Known Activations