INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

ragon

-0.72

asio

-0.69

actly

-0.68

 spores

-0.67

 pickups

-0.67

 flashback

-0.66

anship

-0.66

gdala

-0.66

 Bris

-0.64

anced

-0.64

POSITIVE LOGITS

VB

0.80

swick

0.75

Ru

0.71

VD

0.70

 Ware

0.70

utenberg

0.67

Medic

0.66

leck

0.64

 Murdoch

0.64

 Remain

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.