INDEX

Explanations

references to the act of suicide

oai_token-act-pair · gpt-3.5-turbo

references to suicide

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

ï¸ı

-0.74

lot

-0.71

 Clean

-0.68

eared

-0.67

Cub

-0.67

:\

-0.66

loc

-0.65

ernels

-0.64

Grab

-0.64

Lot

-0.63

POSITIVE LOGITS

 suicide

4.00

 suicides

2.72

 Suicide

2.65

 suicidal

2.24

 homicide

1.62

icide

1.54

 death

1.43

 murder

1.42

 assassination

1.36

 euth

1.35

Activations Density 0.009%

references to the act of suicide

references to suicide

No Comments

No Known Activations