INDEX

Explanations

phrases related to criminal activities involving intent to commit illegal acts

oai_token-act-pair · gpt-3.5-turbo

phrases related to intent to commit a crime or unlawful act

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

 doubling

-0.74

 stressing

-0.70

soDeliveryDate

-0.68

 reminding

-0.67

 iterations

-0.66

 contrasting

-0.66

 anchors

-0.65

 HUGE

-0.65

 looming

-0.64

 enthusi

-0.64

POSITIVE LOGITS

 violate

1.36

 perform

1.31

 deceive

1.30

 injure

1.29

 engage

1.27

 evade

1.26

 commit

1.24

 obtain

1.22

 manipulate

1.21

 distribute

1.17

Activations Density 0.216%

phrases related to criminal activities involving intent to commit illegal acts

phrases related to intent to commit a crime or unlawful act

No Comments

No Known Activations