INDEX

Explanations

detailed complex discussions with specific emphasis on qualifications or formalities

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

following certain words

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

uxxxx

-0.69

 fusca

-0.68

 surla

-0.63

haikusbot

-0.62

Билгалдахарш

-0.62

ValueStyle

-0.62

 ſei

-0.60

LookAnd

-0.56

⪜

-0.55

 beſte

-0.54

POSITIVE LOGITS

SuspendLayout

0.43

 Italijani

0.42

المشاركات

0.36

🛳

0.31

izante

0.31

Únete

0.31

 techniczne

0.30

 demek

0.29

 alguna

0.28

endmodule

0.27

Activations Density 0.139%

detailed complex discussions with specific emphasis on qualifications or formalities

following certain words

No Comments

No Known Activations