INDEX

Explanations

words related to chapters or sections of a text

oai_token-act-pair · gpt-3.5-turbo

references to chapters in a book or report

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

zzle

-0.67

oggles

-0.67

pes

-0.66

 fitt

-0.65

 blaster

-0.64

fitting

-0.63

vant

-0.62

agle

-0.61

ple

-0.59

ws

-0.59

POSITIVE LOGITS

 chapter

1.29

 chapters

1.13

ĸļ

1.07

 Chapter

1.03

chapter

1.02

 Chapters

0.98

Chapter

0.96

 GOODMAN

0.81

apeake

0.79

CHAPTER

0.78

Activations Density 0.007%

words related to chapters or sections of a text

references to chapters in a book or report

No Comments

No Known Activations