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

pondre

-0.46

quiao

-0.41

 DEATH

-0.41

ormais

-0.41

 distribuidor

-0.39

 samh

-0.39

 touristes

-0.38

initComponents

-0.38

 migrator

-0.38

 death

-0.38

POSITIVE LOGITS

:✨

0.96

uxxxx

0.77

 autorytatywna

0.73

httphttps

0.71

 Италијани

0.67

StructEnd

0.64

 resourceCulture

0.64

 noDo

0.63

 typelib

0.62

MessageOf

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.