INDEX

Explanations

phrases with the word "OK" followed by a number

oai_token-act-pair · gpt-3.5-turbo

instances of the word "OK" or variations of it

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

fem

-0.70

 bene

-0.67

 blood

-0.65

 chronic

-0.63

 features

-0.63

 Blood

-0.62

fil

-0.60

ce

-0.60

 weekend

-0.59

mar

-0.59

POSITIVE LOGITS

OK

4.37

OK

2.00

Okay

1.96

Ok

1.92

ok

1.82

Alright

1.53

 Okay

1.44

ok

1.35

OH

1.33

 Alright

1.18

Activations Density 0.013%

phrases with the word "OK" followed by a number

instances of the word "OK" or variations of it

No Comments

No Known Activations