INDEX

Explanations

terms related to gender, specifically the mentions of male and female

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

New Auto-Interp

Configuration

ctigges/pythia-70m-deduped__res-sm_processed/2-res-sm

Prompts (Dashboard)

32,768 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

32,768

Data Type

torch.float32

Hook Name

blocks.2.hook_resid_post

Hook Layer

Architecture

standard

Context Size

128

Dataset

EleutherAI/the_pile_deduplicated

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 applicable

-1.74

 happening

-1.41

ters

-1.39

inson

-1.37

olen

-1.35

 metast

-1.30

Ã¡z

-1.29

 tighter

-1.29

 stolen

-1.27

subseteq

-1.26

POSITIVE LOGITS

¢

2.19

¬

2.12

ļ

2.12

¾

1.98

į

1.96

ĻĤ

1.96

ī

1.90

ģ

1.88

ľ

1.82

ĩ

1.81

Activations Density 0.091%

terms related to gender, specifically the mentions of male and female

No Comments

No Known Activations