INDEX

Explanations

mentions of individuals being arrested

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

natureconservancy

-0.87

 Remastered

-0.81

surv

-0.80

sylv

-0.73

fortune

-0.72

optim

-0.69

 survives

-0.69

interrupted

-0.68

 extinct

-0.68

 Surviv

-0.68

POSITIVE LOGITS

 arra

1.15

 suspicion

1.07

 extradition

1.03

 felony

1.02

DUI

1.00

 Charges

0.99

 misdemeanor

0.96

 indictment

0.92

 handcuffs

0.92

 Arrest

0.91

Activations Density 0.320%

mentions of individuals being arrested

No Comments

No Known Activations