INDEX

Explanations

mentions of individuals holding positions of authority or leadership

oai_token-act-pair · gpt-3.5-turbo

references to individuals in positions of authority or leadership

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

asca

-0.75

uder

-0.73

Ø§ÙĦ

-0.67

aver

-0.65

tnc

-0.64

eros

-0.64

 IMAGES

-0.63

ENC

-0.63

veh

-0.62

Sharp

-0.61

POSITIVE LOGITS

IAL

1.03

ority

1.01

 chairman

0.93

 chair

0.83

 Chairman

0.81

doms

0.79

pins

0.78

ially

0.78

 emer

0.77

 rabbi

0.76

Activations Density 0.013%

mentions of individuals holding positions of authority or leadership

references to individuals in positions of authority or leadership

No Comments

No Known Activations