INDEX

Explanations

references to categories or classifications of items or entities

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

other

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

 widers

-0.50

setupUi

-0.47

 jadx

-0.46

 TestBed

-0.45

 GenerationType

-0.45

principalColumn

-0.44

Discografia

-0.44

KommentareTeilen

-0.44

 yür

-0.44

cury

-0.43

POSITIVE LOGITS

 Other

0.63

Other

0.61

 Autres

0.59

 OTHER

0.57

 Otros

0.56

other

0.56

OTHER

0.54

 Others

0.54

 miscellaneous

0.54

其他

0.52

Activations Density 0.019%

references to categories or classifications of items or entities

other

No Comments

No Known Activations

references to categories or classifications of items or entities

other

No Comments

No Known Activations