INDEX

Explanations

specific legal charges and counts associated with crimes

oai_token-act-pair · gpt-3.5-turbo

phrases related to legal charges or criminal offenses

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

eday

-0.81

uristic

-0.80

days

-0.74

uably

-0.73

soDeliveryDate

-0.72

eworks

-0.72

Tokens

-0.71

 bowls

-0.71

ometers

-0.70

eus

-0.69

POSITIVE LOGITS

 aggravated

1.02

 perjury

1.02

 vandalism

1.01

 extortion

0.95

 Discrimination

0.95

 manslaughter

0.95

 obstruction

0.94

 misdem

0.92

 felony

0.92

 arson

0.90

Activations Density 0.092%

specific legal charges and counts associated with crimes

phrases related to legal charges or criminal offenses

No Comments

No Known Activations