INDEX

Explanations

mentions of particular states or locations

oai_token-act-pair · gpt-3.5-turbo

the word "the" in various contexts throughout the document

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

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

tons

-0.68

packages

-0.68

artifacts

-0.65

forces

-0.65

pee

-0.64

stuff

-0.62

.—

-0.61

lift

-0.60

ansas

-0.59

pees

-0.58

POSITIVE LOGITS

 outset

1.34

 moment

1.21

 same

1.18

 behest

1.14

end

1.14

 beginning

1.03

 forefront

0.99

 conclusion

0.99

 height

0.92

 expense

0.91

Activations Density 0.055%

mentions of particular states or locations

the word "the" in various contexts throughout the document

No Comments

No Known Activations