INDEX

Explanations

references to animated characters and mascots

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

game characters

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

 ویکی‌پدی

-0.63

zzleHttp

-0.61

 autorytatywna

-0.61

CloseOperation

-0.60

 transfieras

-0.58

 الحره

-0.57

باردا

-0.57

 Signalez

-0.57

ंदीखरीदारी

-0.56

RUnlock

-0.56

POSITIVE LOGITS

 mascot

0.59

 mascota

0.42

 cartoon

0.41

ヌーピー

0.38

 cute

0.37

 characters

0.36

icolo

0.34

 캐릭터

0.34

 cartoons

0.34

characters

0.33

Activations Density 0.007%

references to animated characters and mascots

game characters

No Comments

No Known Activations