INDEX

Explanations

names related to titles or positions of power, especially within religious or hierarchical contexts

oai_token-act-pair · gpt-3.5-turbo

terms related to patriarchal structures and authority figures

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

llo

-0.88

lli

-0.87

hig

-0.83

avy

-0.83

cloth

-0.79

llor

-0.79

uten

-0.77

geist

-0.76

lde

-0.74

ppel

-0.74

POSITIVE LOGITS

 patriarch

1.32

 Patriarch

1.06

 patriarchal

0.84

 elders

0.74

 ç¥ŀ

0.72

 illegitimate

0.71

owship

0.69

 Torah

0.68

 Klaus

0.68

 patriarchy

0.67

Activations Density 0.023%

names related to titles or positions of power, especially within religious or hierarchical contexts

terms related to patriarchal structures and authority figures

No Comments

No Known Activations