INDEX

Explanations

numeric sequences that are not in a particular pattern

oai_token-act-pair · gpt-3.5-turbo

unusual characters or symbols in text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

 Polk

-0.90

 Chocolate

-0.84

 Meter

-0.81

 Mobil

-0.81

 Massive

-0.81

 Pages

-0.79

 Meteor

-0.79

Nep

-0.79

 Coffee

-0.78

 Bacon

-0.78

POSITIVE LOGITS

subject

1.30

average

1.24

since

1.22

when

1.21

there

1.21

sufficient

1.20

because

1.19

shared

1.18

thus

1.18

total

1.18

Activations Density 0.286%

numeric sequences that are not in a particular pattern

unusual characters or symbols in text

No Comments

No Known Activations