INDEX

Explanations

words related to fraudulent activities

oai_token-act-pair · gpt-3.5-turbo

references to fraudulent activities

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

%"

-0.72

%:

-0.67

Own

-0.67

lihood

-0.64

Dak

-0.64

joy

-0.62

 LONG

-0.60

 Roses

-0.60

aple

-0.59

VILLE

-0.59

POSITIVE LOGITS

 fraud

1.20

ulence

1.17

ulent

1.05

 scam

0.92

 scams

0.85

sters

0.80

ster

0.80

ously

0.79

 perpetrated

0.79

 Fraud

0.79

Activations Density 0.015%

words related to fraudulent activities

references to fraudulent activities

No Comments

No Known Activations