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

 therap

-0.73

 challeng

-0.72

 COUR

-0.64

 Citiz

-0.63

 Nurs

-0.61

 Dragonbound

-0.61

puter

-0.60

essor

-0.60

 sidx

-0.58

 Afric

-0.58

POSITIVE LOGITS

IJ

0.68

ingen

0.67

 result

0.63

 torpedo

0.62

itably

0.61

STON

0.59

cerning

0.59

 eclipse

0.59

usha

0.59

 shade

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.