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

pse

-0.74

 constitu

-0.74

 confir

-0.71

 precedence

-0.70

govtrack

-0.66

aca

-0.64

 carbohyd

-0.64

 Heidi

-0.63

 revelations

-0.62

srf

-0.62

POSITIVE LOGITS

Gall

0.72

ripp

0.68

ships

0.68

©¶æ¥µ

0.68

xi

0.65

 Tycoon

0.65

System

0.64

rill

0.62

 Jaguar

0.62

 harmless

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.