INDEX

Explanations

legal terms and criminal charges

oai_token-act-pair · gpt-3.5-turbo

legal terminology related to degrees of murder and other criminal charges

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 Leader

-0.68

Sov

-0.67

ãĤ£

-0.66

 valleys

-0.64

 Slim

-0.63

 Stud

-0.62

 Patriarch

-0.62

 watered

-0.60

meric

-0.59

eus

-0.59

POSITIVE LOGITS

degree

1.55

 degree

1.26

 Degree

1.26

 felony

1.16

 burglary

1.04

 manslaughter

0.99

 homicide

0.99

 misdemeanor

0.98

 aggravated

0.94

 offenses

0.94

Activations Density 0.052%

legal terms and criminal charges

legal terminology related to degrees of murder and other criminal charges

No Comments

No Known Activations