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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

usammen

-0.36

say

-0.35

inį

-0.34

stringValue

-0.34

Климат

-0.33

fly

-0.33

maxlen

-0.33

retung

-0.33

Mc

-0.32

 olsa

-0.32

POSITIVE LOGITS

Datuak

2.75

Datos

2.14

 Datos

1.44

GEBURTSDATUM

1.22

Portail

1.18

بوابة

1.11

 datos

1.04

 DATOS

1.03

datos

0.97

 autorytatywna

0.95

Activations Density 0.000%

No Known Activations

This feature has no known activations.