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

undai

-0.86

sburgh

-0.84

ottesville

-0.82

vertising

-0.78

 traged

-0.74

ersen

-0.72

ammad

-0.68

espie

-0.68

kinson

-0.67

GPU

-0.67

POSITIVE LOGITS

¼

0.73

º

0.71

CDC

0.69

 nutshell

0.67

onet

0.67

ĳ

0.67

 COUR

0.66

 Parables

0.66

etary

0.65

cial

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.