INDEX

Explanations

code or programming-related syntax elements

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

code, markup, scientific notation

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

 المعيارى

-0.71

 disponibilités

-0.53

 AssemblyCulture

-0.52

ulemon

-0.46

 Riproduzione

-0.44

eseorang

-0.43

Rüyada

-0.42

цезда

-0.40

InputBorder

-0.40

EnableWeb

-0.40

POSITIVE LOGITS

%%

0.74

ViewFeatures

0.69

endphp

0.57



0.56

uxxxx

0.56

Bakgrunnsstoff

0.54

%%

0.51

endregion

0.50

 Denn

0.49

 CURIAM

0.48

Activations Density 0.002%

code or programming-related syntax elements

code, markup, scientific notation

No Comments

No Known Activations