INDEX

Explanations

questions

oai_token-act-pair · gpt-3.5-turbo

questions, particularly those that provoke thought or inquiry

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

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

apan

-0.77

 transition

-0.76

éĹĺ

-0.74

 insertion

-0.72

agne

-0.72

 fabrication

-0.69

ense

-0.69

hesda

-0.68

roup

-0.66

 practition

-0.66

POSITIVE LOGITS

 Nope

1.53

 Absolutely

1.52

 Possibly

1.44

 Probably

1.42

 Certainly

1.34

Yes

1.33

 Answer

1.28

YES

1.24

 Would

1.23

 Sure

1.20

Activations Density 0.102%

questions

questions, particularly those that provoke thought or inquiry

No Comments

No Known Activations