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

bcryptjs

-0.71



-0.66

 CreateTagHelper

-0.65

TagMode

-0.62

principalTable

-0.61

 propOrder

-0.59

 transfieras

-0.56

 queſta

-0.54

 Wicidata

-0.54

EDEFAULT

-0.54

POSITIVE LOGITS

Trọng

0.40

 pres

0.36

Gew

0.36

BARA

0.36

fläche

0.36

0.35

EditorBrowsable

0.34

림

0.34

FORMANCE

0.34

ges

0.34

Activations Density 0.000%

No Known Activations

This feature has no known activations.