INDEX

Explanations

terms related to intricate pathways and structures, particularly mazes and labyrinths

oai_token-act-pair · gpt-4o-mini Triggered by @bot

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

PhysRev

-0.45

Officer

-0.43

officer

-0.42

ọn

-0.41

 Britton

-0.40

Philli

-0.39

AddWithValue

-0.39

('

-0.38

boxylic

-0.37

 Cummins

-0.37

POSITIVE LOGITS

 Maze

1.38

 maze

1.35

Maze

1.12

maze

1.10

 labyrinth

1.05

 Labyrinth

1.02

abyrinth

0.96

 laby

0.94

 aprendizagem

0.69

BagLayout

0.63

Activations Density 0.004%

terms related to intricate pathways and structures, particularly mazes and labyrinths

No Comments

No Known Activations