INDEX

Explanations

phrases with specific patterns of characters, possibly indicating a formatting style or code structure

oai_token-act-pair · gpt-3.5-turbo

instances of emphasis or punctuation in the text

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

 credential

-0.79

 neighb

-0.78

 jailed

-0.77

 seized

-0.77

 stunt

-0.74

 increasingly

-0.74

 dogged

-0.73

 outraged

-0.73

 crunch

-0.73

 likewise

-0.73

POSITIVE LOGITS

Introduction

1.51

Overview

1.36

Description

1.34

Hello

1.32

FAQ

1.31

Installation

1.28

Disclaimer

1.27

âķ

1.26

Quote

1.26

Features

1.26

Activations Density 0.146%

phrases with specific patterns of characters, possibly indicating a formatting style or code structure

instances of emphasis or punctuation in the text

No Comments

No Known Activations