INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

psons

-0.84

alle

-0.71

ARC

-0.70

alian

-0.69

APD

-0.68

ogens

-0.68

ileaks

-0.67

KR

-0.66

aido

-0.66

sighted

-0.66

POSITIVE LOGITS

="#

0.79

rou

0.70

ominium

0.69

iton

0.69

 Gamble

0.67

 mutation

0.62

 Pinball

0.62

NBN

0.62

 Booth

0.62

 discounted

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.