INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

âĶĢâĶĢâĶĢâĶĢâĶĢâĶĢâĶĢâĶĢ

-0.79

Side

-0.72

 Radiant

-0.70

 Confederation

-0.70

 Spiral

-0.65

Chel

-0.65

âĶĢâĶĢâĶĢâĶĢ

-0.65

 Colossus

-0.64

 Hive

-0.63

ãĤ¼ãĤ¦ãĤ¹

-0.62

POSITIVE LOGITS

rates

0.78

fu

0.76

itle

0.70

illac

0.68

odka

0.67

doms

0.67

ptin

0.67

gui

0.65

hip

0.64

gerald

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.