INDEX

Explanations

references to the "Harry Potter" series

oai_token-act-pair · gpt-3.5-turbo

references to the Harry Potter franchise

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

oples

-0.72

AY

-0.67

¡

-0.66

CHR

-0.64

 condensed

-0.63

 prevailing

-0.62

Ĩ

-0.62

PER

-0.62

ĻĤ

-0.61

resp

-0.60

POSITIVE LOGITS

 Potter

1.49

 Weasley

1.12

uggle

1.04

 Malfoy

0.94

 Granger

0.92

more

0.86

atoes

0.86

 Rowling

0.84

fleet

0.80

Pot

0.80

Activations Density 0.007%

references to the "Harry Potter" series

references to the Harry Potter franchise

No Comments

No Known Activations