INDEX

Explanations

academic references in a specific format with titles, authors, dates, and paper themes

oai_token-act-pair · gpt-3.5-turbo

references to academic papers and their citations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 rented

-0.80

 cooked

-0.79

 wandered

-0.77

 champ

-0.76

psons

-0.75

 anytime

-0.73

 departed

-0.73

 taco

-0.73

 snatched

-0.73

 entertain

-0.72

POSITIVE LOGITS

 Definitions

1.34

 Factors

1.32

 Comparison

1.31

 Understanding

1.31

 Principles

1.30

Understanding

1.30

 Evaluation

1.29

 Effects

1.29

 Influence

1.27

Introduction

1.26

Activations Density 0.163%

academic references in a specific format with titles, authors, dates, and paper themes

references to academic papers and their citations

No Comments

No Known Activations