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

 république

-0.39

de

-0.35

↵

-0.33

ts

-0.32

 piani

-0.32

en

-0.31

（

-0.31

-0.30

 terres

-0.30

le

-0.30

POSITIVE LOGITS

Хьажоргаш

0.88

OGND

0.84

 Numerade

0.83

 kasarigan

0.82

хьтан

0.81

 Италијани

0.77

 незавершена

0.77

 nakalista

0.77

تقاوى

0.77

 للمعارف

0.76

Activations Density 0.000%

No Known Activations

This feature has no known activations.