INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 9-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_9/width_131k/average_l0_121

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.9.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

 Metropolitan

-0.32

之

-0.30

 getattr

-0.29

mè

-0.29

 Rauch

-0.28

<eos>

-0.28

-0.27

 usus

-0.27

 remo

-0.26

 lack

-0.25

POSITIVE LOGITS

 createState

7.13

 initComponents

1.44

setVerticalGroup

1.23

 createSlice

1.07

 MonoBehaviour

1.03

MemoryWarning

0.97

StoryboardSegue

0.95

 StatelessWidget

0.94

httphttps

0.93

SharedCtor

0.90

Activations Density 0.000%

No Known Activations

This feature has no known activations.