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

othe

-0.75

Ü

-0.73

ergy

-0.73

GV

-0.68

Els

-0.63

gey

-0.62

alon

-0.62

uggle

-0.62

itaire

-0.60

aunch

-0.60

POSITIVE LOGITS

 Instead

0.75

Nug

0.74

Ĩ

0.73

 Lomb

0.73

ÄŁ

0.72

 Pepe

0.68

 Pione

0.67

 Krug

0.66

Laf

0.66

borg

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.