INDEX

Explanations

proper nouns related to literature

oai_token-act-pair · gpt-3.5-turbo

mentions of a specific person or entity, likely related to identity or prominence

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 calib

-0.76

 marqu

-0.68

Met

-0.68

 cent

-0.67

 diam

-0.63

 central

-0.61

 Worlds

-0.61

 quart

-0.58

Europe

-0.58

 Robots

-0.58

POSITIVE LOGITS

hee

4.82

hey

1.39

hea

1.37

hend

1.24

heon

1.22

he

1.21

het

1.19

heet

1.18

kee

1.13

heed

1.11

Activations Density 0.010%

proper nouns related to literature

mentions of a specific person or entity, likely related to identity or prominence

No Comments

No Known Activations