INDEX

Explanations

photographs with corresponding credits

oai_token-act-pair · gpt-3.5-turbo

references to photos

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

amped

-0.70

amping

-0.67

ecause

-0.61

outine

-0.60

 footing

-0.60

uckle

-0.59

 judgement

-0.57

osc

-0.56

halla

-0.55

EEK

-0.55

POSITIVE LOGITS

journal

1.11

 illustration

1.08

 courtesy

1.04

 gallery

1.00

 Gallery

0.98

 Illust

0.95

 Credit

0.93

 Courtesy

0.89

 caption

0.88

 credit

0.86

Activations Density 0.017%

photographs with corresponding credits

references to photos

No Comments

No Known Activations