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

ische

-0.77

chell

-0.74

ICS

-0.72

former

-0.69

LR

-0.69

lite

-0.66

essional

-0.66

lr

-0.66

ERN

-0.65

 Omni

-0.65

POSITIVE LOGITS

ayed

0.81

 seiz

0.77

checked

0.75

utherford

0.69

oice

0.68

ãģĨ

0.68

ata

0.67

assed

0.67

ottesville

0.67

 Class

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.