INDEX

Explanations

specific mentions of the character "Harry Potter" in various contexts

oai_token-act-pair · gpt-3.5-turbo

references to the Harry Potter series

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

asio

-0.75

©¶æ

-0.75

oples

-0.74

Ĩ

-0.68

resp

-0.68

¡

-0.68

rikes

-0.67

§

-0.66

rontal

-0.64

ĻĤ

-0.64

POSITIVE LOGITS

 Potter

1.17

 Weasley

0.91

haus

0.83

 Granger

0.83

 Gand

0.81

 Dumbledore

0.81

 Rowling

0.79

more

0.78

atoes

0.76

wra

0.76

Activations Density 0.013%

specific mentions of the character "Harry Potter" in various contexts

references to the Harry Potter series

No Comments

No Known Activations