INDEX

Explanations

words related to authority figures, especially bosses

oai_token-act-pair · gpt-3.5-turbo

mentions of the word "boss."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

Ô

-0.76

dn

-0.69

 Scholars

-0.68

 Hampton

-0.64

 libraries

-0.62

years

-0.61

 Recogn

-0.59

Footnote

-0.59

 Libraries

-0.59

 evid

-0.58

POSITIVE LOGITS

 boss

3.93

 bosses

2.96

boss

2.69

Boss

2.40

 Boss

2.23

 supervisor

1.51

 manager

1.35

 employer

1.30

 superiors

1.27

 mistress

1.23

Activations Density 0.015%

words related to authority figures, especially bosses

mentions of the word "boss."

No Comments

No Known Activations