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

 assassinate

-0.82

 hither

-0.81

eatures

-0.75

ategory

-0.75

asta

-0.75

 seiz

-0.74

Reviewer

-0.73

ntil

-0.72

pection

-0.71

 conclud

-0.70

POSITIVE LOGITS

 Group

1.19

 Bentley

0.68

 GeForce

0.68

framework

0.68

wick

0.65

 GROUP

0.64

 batteries

0.63

 Holdings

0.62

 Jaguar

0.62

ERN

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.