INDEX

Explanations

images or content credits

oai_token-act-pair · gpt-3.5-turbo

references to credit attribution in images or content

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

execute

-0.66

aiman

-0.66

ernels

-0.62

 Sell

-0.62

 TAMADRA

-0.61

rowth

-0.61

xit

-0.60

 invested

-0.59

rap

-0.58

affected

-0.58

POSITIVE LOGITS

 courtesy

3.87

tesy

2.18

 Courtesy

1.96

cour

1.69

Courtesy

1.52

 thanks

1.34

 compliments

1.29

 decency

1.14

thanks

1.11

 Â©

1.00

Activations Density 0.018%

images or content credits

references to credit attribution in images or content

No Comments

No Known Activations