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

↵

-0.47

 Schreiber

-0.44

↵↵

-0.42

list

-0.41

 Ziegler

-0.41

 Kombin

-0.40

 Sommers

-0.40

相

-0.39

text

-0.39

-0.38

POSITIVE LOGITS

parsedMessage

0.93

出版年

0.82

 незавершена

0.82

tagHelperRunner

0.82

httphttps

0.80



0.75

 Taktlose

0.72

Personensuche

0.72

المناصب

0.71

AddTagHelper

0.70

Activations Density 0.000%

No Known Activations

This feature has no known activations.