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

новниш

-0.84

帖最后由

-0.60

stdc

-0.56

afficheront

-0.56

 Infórmanos

-0.53

tanleria

-0.52

OGND

-0.52

CloseOperation

-0.52

kloped

-0.51

 propOrder

-0.51

POSITIVE LOGITS

0.78

 Jambi

0.38

Datetime

0.36

aarrggbb

0.34

pæ

0.34

Technik

0.34

 fresca

0.34

Darm

0.33

 stdin

0.33

 corrom

0.33

Activations Density 0.000%

No Known Activations

This feature has no known activations.