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

 Yong

-0.67

 Maid

-0.67

Maker

-0.66

Op

-0.66

 shin

-0.64

Lu

-0.64

 Living

-0.63

Su

-0.62

Hum

-0.61

 Clicker

-0.61

POSITIVE LOGITS

 States

1.04

apses

0.80

iliate

0.78

ournals

0.75

withstanding

0.74

orse

0.74

zbollah

0.73

pta

0.73

iland

0.72

irection

0.71

Activations Density 0.000%

No Known Activations

This feature has no known activations.