INDEX

Explanations

alphanumeric sequences combined with special characters

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 financing

-0.82

 textbooks

-0.80

 proble

-0.79

 tenants

-0.79

 credential

-0.78

 curric

-0.77

 unfavorable

-0.77

 detainees

-0.77

 indict

-0.77

 UNESCO

-0.75

POSITIVE LOGITS

1.28

Gaming

1.26

Magikarp

1.23

guy

1.21

bear

1.19

1.18

pie

1.17

gamer

1.16

1.15

Activations Density 0.258%

alphanumeric sequences combined with special characters

No Comments

No Known Activations