INDEX

Explanations

quotations enclosed in double quotes

oai_token-act-pair · gpt-3.5-turbo

quotation marks and speech indicators in the text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

 adjud

-0.81

 cram

-0.81

 distribut

-0.78

 sway

-0.78

 derby

-0.75

 dominate

-0.75

 flared

-0.75

 developmental

-0.75

 schedule

-0.74

 favor

-0.73

POSITIVE LOGITS

We

1.32

Absolutely

1.28

1.26

Our

1.24

Whoever

1.23

There

1.22

You

1.19

Everything

1.19

Anything

1.18

Everyone

1.18

Activations Density 0.079%

quotations enclosed in double quotes

quotation marks and speech indicators in the text

No Comments

No Known Activations