INDEX

Explanations

No Explanations Found

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

AndEndTag

-0.52

 Signalez

-0.45

styleUrls

-0.44

 Common

-0.40

SuppressMessage

-0.40

mobileqq

-0.40

AutoScaleMode

-0.39

ExecuteNonQuery

-0.39

MethodImpl

-0.39

 StyleSheet

-0.38

POSITIVE LOGITS

0.67

uvwxyz

0.52

 cív

0.49

 nachfol

0.47

 ujednoznacz

0.47

 táctil

0.47

 médicale

0.46

何より

0.46

Personensuche

0.46

levis

0.46

Activations Density 0.000%

No Known Activations

This feature has no known activations.