INDEX

Explanations

phrases related to ethics, morality, and law

oai_token-act-pair · gpt-3.5-turbo

references to violence or harmful actions against people and groups

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

Typically

-0.76

luster

-0.74

 Typically

-0.72

TPPStreamerBot

-0.70

Wow

-0.70

bright

-0.69

 Glow

-0.69

 Velocity

-0.69

Depth

-0.69

 Nerd

-0.68

POSITIVE LOGITS

 assassination

1.43

 treason

1.42

 genocide

1.42

 assass

1.41

 slander

1.40

 massacres

1.31

 blasphemy

1.31

 persecution

1.30

 blasp

1.26

 assassinate

1.25

Activations Density 0.736%

phrases related to ethics, morality, and law

references to violence or harmful actions against people and groups

No Comments

No Known Activations