INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

chel

-0.74

Contents

-0.73

iatus

-0.72

Father

-0.71

Cra

-0.70

adem

-0.70

udi

-0.69

mens

-0.68

Edited

-0.66

idis

-0.66

POSITIVE LOGITS

ufact

0.78

oeuv

0.73

yip

0.71

 ACTIONS

0.69

oÄŁ

0.68

 baggage

0.68

RAD

0.66

srf

0.64

arching

0.61

phase

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.