INDEX

Explanations

words and phrases related to violent actions or events

oai_token-act-pair · gpt-3.5-turbo

references to violent incidents or behaviors

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

ocular

-0.84

heit

-0.83

obyl

-0.78

printed

-0.78

ovember

-0.76

ļéĨĴ

-0.75

cellent

-0.74

é¾įå

-0.73

STER

-0.73

regon

-0.73

POSITIVE LOGITS

 violence

0.97

 retribution

0.88

 assault

0.88

 violent

0.86

 offenders

0.85

 suppression

0.84

 gang

0.83

 deaths

0.82

 Islamist

0.81

 repression

0.81

Activations Density 0.013%

words and phrases related to violent actions or events

references to violent incidents or behaviors

No Comments

No Known Activations