INDEX

Explanations

positively descriptive phrases about books or written works

oai_token-act-pair · gpt-3.5-turbo

positive adjectives and phrases describing written works

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

-1.17

netflix

-0.80

Xbox

-0.78

ereo

-0.77

 Twitch

-0.76

VK

-0.75

AMI

-0.75

 valve

-0.75

twitch

-0.74

 Tradable

-0.74

POSITIVE LOGITS

 chapters

1.56

 Chapters

1.35

chapter

1.33

 chapter

1.32

 paragraphs

1.19

 appendix

1.18

 essays

1.18

 passages

1.15

 Introduction

1.12

 sections

1.11

Activations Density 0.427%

positively descriptive phrases about books or written works

positive adjectives and phrases describing written works

No Comments

No Known Activations