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

estic

-0.80

achus

-0.78

ifact

-0.78

nesday

-0.76

manship

-0.71

oggle

-0.70

swer

-0.68

igham

-0.68

uggest

-0.66

atl

-0.65

POSITIVE LOGITS

 Void

0.76

humans

0.69

IJ

0.68

Living

0.68

Definition

0.68

eem

0.68

Reviewer

0.67

Ent

0.67

hani

0.67

Contents

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.