INDEX

Explanations

quantitative data related to specific categories or measurements

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

percentages and fractions

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

 AssemblyTitle

-0.50

exitRule

-0.40

 wireType

-0.35

endregion

-0.33

WithMany

-0.32

anskje

-0.31

 îna

-0.31

 blow

-0.29

 hjälp

-0.29

soil

-0.29

POSITIVE LOGITS

 Meksiku

0.69

uxxxx

0.61

AppMethodBeat

0.59

 resourceCulture

0.56

+#+

0.53

ThemeOverlay

0.52

__*/

0.51

 يتيمه

0.50

Médaille

0.50

SequentialGroup

0.50

Activations Density 0.002%

quantitative data related to specific categories or measurements

percentages and fractions

No Comments

No Known Activations