INDEX

Explanations

No Explanations Found

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

bud

-0.80

 favour

-0.78

 favor

-0.73

 hiding

-0.73

 jealous

-0.73

 wire

-0.72

 lying

-0.72

 surrender

-0.71

 pitch

-0.71

 escaping

-0.70

POSITIVE LOGITS

Hist

1.13

Introduction

1.09

Recent

1.05

Grade

1.03

Which

0.99

WORK

0.99

Topics

0.98

Building

0.98

Database

0.97

Visit

0.97

Activations Density 0.000%

No Known Activations

This feature has no known activations.