INDEX

Explanations

words related to tasks or instructions

oai_token-act-pair · gpt-3.5-turbo

actions related to personal responsibility and preparation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

)]

-0.57

translation

-0.56

":""},{"

-0.54

çīĪ

-0.53

Translation

-0.52

TPPStreamerBot

-0.52

Their

-0.51

"]=>

-0.51

andowski

-0.51

Compar

-0.50

POSITIVE LOGITS

 yourself

1.83

 your

1.83

 yourselves

1.52

 YOUR

1.51

your

1.47

 Yourself

1.37

 Your

1.37

Your

1.36

 yours

1.25

you

0.91

Activations Density 0.945%

words related to tasks or instructions

actions related to personal responsibility and preparation

No Comments

No Known Activations