INDEX

Explanations

No Explanations Found

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_131k/average_l0_114

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

덟

-0.35

mbggenerated

-0.30

 ladite

-0.30

 Châ

-0.29

 Neust

-0.29

 renovation

-0.28

gebir

-0.28

 action

-0.28

 Blum

-0.28

the

-0.28

POSITIVE LOGITS

1.09

0.92

 メンテナ

0.90

ſicht

0.89

iſchen

0.89

ſſung

0.88

ðsíða

0.88

 erſt

0.88

majánló

0.88

 パンチラ

0.87

Activations Density 0.000%

No Known Activations

This feature has no known activations.