INDEX

Explanations

references to specific episodes of a TV show called "Dragon Ball Super."

oai_token-act-pair · gpt-3.5-turbo

references to the franchise "Dragon Ball"

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

ALS

-0.76

 Trinidad

-0.72

ablished

-0.71

 Crosby

-0.70

 Morse

-0.69

chell

-0.68

 Beckham

-0.68

 Notting

-0.67

oci

-0.67

EMS

-0.67

POSITIVE LOGITS

fly

1.28

flies

1.17

hawk

0.98

bone

0.95

Fly

0.90

flight

0.89

claw

0.89

lance

0.88

plate

0.81

 dragons

0.81

Activations Density 0.032%

references to specific episodes of a TV show called "Dragon Ball Super."

references to the franchise "Dragon Ball"

No Known Activations