INDEX

Explanations

quotations, particularly those emphasizing strong opinions or beliefs

oai_token-act-pair · gpt-3.5-turbo

quotation marks and dialogue

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

 Armenian

-0.80

 Closing

-0.72

 Ange

-0.71

 tabloid

-0.71

ERY

-0.71

 daylight

-0.70

 consequential

-0.69

 Daylight

-0.69

 Anthropology

-0.69

 constructive

-0.68

POSITIVE LOGITS

could

1.61

would

1.56

should

1.51

had

1.49

didn

1.48

doesn

1.43

did

1.42

saw

1.40

must

1.38

might

1.38

Activations Density 0.136%

quotations, particularly those emphasizing strong opinions or beliefs

quotation marks and dialogue

No Comments

No Known Activations