INDEX

Explanations

commands or instructions ending with a period, typically used in step-by-step guides

oai_token-act-pair · gpt-3.5-turbo

sentence-ending punctuation marks

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 indebted

-0.80

 elusive

-0.80

 dared

-0.80

 affili

-0.78

 boycot

-0.73

 plague

-0.73

 choking

-0.72

 unexplained

-0.71

 exotic

-0.70

 antit

-0.70

POSITIVE LOGITS

 Then

1.74

 Alternatively

1.74

 Afterwards

1.66

 Otherwise

1.55

 Depending

1.55

 Ideally

1.48

 Each

1.46

 Additionally

1.44

 However

1.39

 Assuming

1.38

Activations Density 0.303%

commands or instructions ending with a period, typically used in step-by-step guides

sentence-ending punctuation marks

No Comments

No Known Activations