INDEX

Explanations

references to religious worship and practices

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 25-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_25/width_16k/average_l0_116

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.25.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

AccessorTable

-0.89

Rhestr

-0.64

FZ

-0.57

lamabad

-0.56

setInput

-0.51

India

-0.49

codable

-0.49

 DELHI

-0.49

 India

-0.49

 Islamabad

-0.49

POSITIVE LOGITS

 Lord

0.88

Lord

0.75

 Goddess

0.72

 devotees

0.71

 deity

0.70

 lord

0.68

 idol

0.67

Goddess

0.64

 Deity

0.64

 deities

0.63

Activations Density 0.141%

references to religious worship and practices

No Comments

No Known Activations