INDEX

Explanations

significant phrases in the form of special characters and numbers

oai_token-act-pair · gpt-3.5-turbo

end-of-document markers

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

enegger

-0.83

Peb

-0.80

 scattering

-0.76

 princ

-0.73

 scatter

-0.68

anwhile

-0.67

 predec

-0.65

 Afric

-0.65

Tid

-0.65

gad

-0.64

POSITIVE LOGITS

º

1.40

¹

1.25

£

1.19

¬

1.18

Į

1.14

į

1.14

§

1.13

Ĵ

1.11

ı

1.08

ĭ

1.06

Activations Density 0.286%

significant phrases in the form of special characters and numbers

end-of-document markers

No Comments

No Known Activations