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

 ComVisible

-0.85

PyExc

-0.84

oa̍t

-0.81

 betweenstory

-0.80

 OFDb

-0.78

AsUp

-0.78

 パンチラ

-0.77

complexContent

-0.76

RenderAtEndOf

-0.76

 témoig

-0.75

POSITIVE LOGITS

Initializable

0.89

 table

0.31

 diesmal

0.29

 nobre

0.26

 juges

0.25

Table

0.25

 vrais

0.25

書の

0.24

 recursos

0.24

 milliers

0.23

Activations Density 0.000%

No Known Activations

This feature has no known activations.