INDEX

Explanations

versions of software or code snippets with specific identifiers

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

ĪĴ

-0.83

 duplication

-0.61

 Ellison

-0.61

--------------------------------------------------------

-0.60

 Defenders

-0.60

éĥ

-0.60

 Flavoring

-0.59

 juggling

-0.58

 contrad

-0.58

ãĤ³

-0.58

POSITIVE LOGITS

irgin

1.28

olution

1.23

irus

1.21

iolet

1.20

ortex

1.17

intage

1.15

endor

1.12

ampire

1.11

apor

1.10

ascular

1.09

Activations Density 0.043%

versions of software or code snippets with specific identifiers

No Comments

No Known Activations