INDEX

Explanations

phrases related to politics and public figures

oai_token-act-pair · gpt-3.5-turbo

instances of criticism or negative commentary directed at political figures or organizations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

"}],"

-0.85

 versa

-0.78

obbies

-0.74

ogether

-0.74

 guiActiveUn

-0.73

ãĤ©

-0.72

è¦ļéĨĴ

-0.71

cellaneous

-0.70

EStream

-0.69

oufl

-0.69

POSITIVE LOGITS

 Donald

1.25

 Hillary

1.10

 Bernie

1.09

 Kanye

1.04

 SpaceX

1.04

 UKIP

1.01

 President

1.00

 Narendra

0.99

 Conor

0.97

 Barack

0.97

Activations Density 0.605%

phrases related to politics and public figures

instances of criticism or negative commentary directed at political figures or organizations

No Comments

No Known Activations