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

 Others

-0.46

 Vikipedi

-0.43

Others

-0.40

まま

-0.39

 яко

-0.36

czo

-0.35

 noqa

-0.35

///</

-0.35

Assets

-0.34

 Troy

-0.34

POSITIVE LOGITS

بینی

2.83

 بینی

0.99

‌شده

0.93

 مرئيه

0.69

Hozzáférés

0.65

Consulta

0.65

tagHelperRunner

0.64

 CreateTagHelper

0.64

 consultato

0.61

Portály

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.