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

 gangs

-0.49

 gang

-0.41

ColumnHeaders

-0.41

 bubbles

-0.39

ymce

-0.39

Viki

-0.39

 gods

-0.38

 variety

-0.38

よう

-0.38

 вида

-0.38

POSITIVE LOGITS

IActionResult

5.25

 IActionResult

1.79

ActionResult

1.37

 ActionResult

1.23

 HttpNotFound

1.16

migrationBuilder

0.79

StoryboardSegue

0.74

 ControllerBase

0.74

AntiForgeryToken

0.73

 ModelExpression

0.71

Activations Density 0.000%

No Known Activations

This feature has no known activations.