INDEX

Explanations

mentions of the word "anime"

oai_token-act-pair · gpt-3.5-turbo

references to anime

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 Lazarus

-0.78

 Prepar

-0.73

GM

-0.70

 Mour

-0.69

mate

-0.68

Tib

-0.67

 McKay

-0.67

Tah

-0.66

 Painter

-0.66

emaker

-0.66

POSITIVE LOGITS

 anime

2.79

 Anime

2.25

 ISPs

1.22

 Comcast

1.22

 manga

1.16

 NASCAR

1.11

 Netflix

1.10

 comics

1.04

Netflix

1.03

 cartoons

1.02

Activations Density 0.029%

mentions of the word "anime"

references to anime

No Comments

No Known Activations