INDEX

Explanations

mentions of individuals holding positions of authority or leadership

oai_token-act-pair · gpt-3.5-turbo

references to individuals holding the title of "chair" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

 IMAGES

-0.81

 tatt

-0.65

 tolerance

-0.65

 suscept

-0.65

 kindness

-0.63

 Ultron

-0.62

 Apocalypse

-0.61

 conformity

-0.61

 DATA

-0.60

 righteousness

-0.60

POSITIVE LOGITS

woman

1.35

person

1.30

manship

1.27

lift

1.14

women

1.12

mans

1.11

men

1.10

chair

1.10

holder

1.06

pers

1.00

Activations Density 0.026%

mentions of individuals holding positions of authority or leadership

references to individuals holding the title of "chair" in various contexts

No Comments

No Known Activations