INDEX

Explanations

terms and phrases relating to principles and values, particularly in a moral or ethical context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

 يتيمه

-0.69

قایناقلار

-0.69

oredCriteria

-0.67

 houſe

-0.64

expandindo

-0.63

verwijspagina

-0.63

RRect

-0.61

 estekak

-0.60

adaptiveStyles

-0.59

RenderAtEndOf

-0.59

POSITIVE LOGITS

 principles

1.63

principles

1.41

 principle

1.38

 Principles

1.37

 philosophy

1.33

 tenets

1.29

Principles

1.29

 princípios

1.28

 values

1.24

principle

1.21

Activations Density 0.330%

terms and phrases relating to principles and values, particularly in a moral or ethical context

No Comments

No Known Activations