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

+:+

-0.47

GEBURTS

-0.45

 NSCoder

-0.44

ljiv

-0.44

 pitié

-0.44

แก่

-0.43

 adota

-0.43

 Bezir

-0.43

 Sünde

-0.43

 coklat

-0.42

POSITIVE LOGITS

complexContent

0.63

TypedDataSet

0.57

migrationBuilder

0.56

 صوتيه

0.55

ViewFeatures

0.54

 autorytatywna

0.49

 initComponents

0.48

AccessorTable

0.47

ISupport

0.47

WebElementEntity

0.47

Activations Density 0.000%

No Known Activations

This feature has no known activations.