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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Denver

-0.73

bet

-0.67

Nap

-0.66

Boston

-0.64

 binge

-0.62

 mornings

-0.62

shot

-0.61

ividual

-0.61

comm

-0.61

Ø¹

-0.59

POSITIVE LOGITS

↵

0.85

uras

0.79

aga

0.77

 Cooldown

0.75

↵↵

0.75

rity

0.72

areth

0.66

 Reeves

0.65

 Written

0.65

uran

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.