INDEX

Explanations

references to historical events, particularly focusing on wars and political figures from the past

oai_token-act-pair · gpt-3.5-turbo

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

 Fukushima

-0.93

 meltdown

-0.87

externalActionCode

-0.84

ÄŁ

-0.83

Shin

-0.83

 robotic

-0.82

zona

-0.82

zbollah

-0.81

achev

-0.79

 hijab

-0.78

POSITIVE LOGITS

1.28

 colonists

1.24

 Crom

1.23

 Scots

1.22

 Cornwall

1.22

eenth

1.20

1.17

 Carth

1.17

1.16

Activations Density 0.675%

references to historical events, particularly focusing on wars and political figures from the past

No Comments

No Known Activations