INDEX

Explanations

references to theft or stolen items

oai_token-act-pair · gpt-3.5-turbo

references to stolen items

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ichick

-0.83

arya

-0.80

arist

-0.77

shr

-0.73

guy

-0.73

amus

-0.72

ingo

-0.71

upon

-0.71

clinical

-0.70

abe

-0.69

POSITIVE LOGITS

 stolen

1.30

 thefts

0.99

 Theft

0.97

 stealing

0.96

 theft

0.86

 steal

0.84

 stole

0.83

 thief

0.83

 thieves

0.81

 snatched

0.77

Activations Density 0.010%

references to theft or stolen items

references to stolen items

No Comments

No Known Activations