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

 Franch

-0.71

Gir

-0.68

MPH

-0.63

Tyr

-0.63

RR

-0.61

ogeneous

-0.60

etus

-0.60

Sor

-0.59

Coh

-0.58

rika

-0.58

POSITIVE LOGITS

pict

0.78

bugs

0.78

photos

0.77

achev

0.74

thumbnails

0.74

DragonMagazine

0.74

EMS

0.74

ube

0.72

imb

0.69

complex

0.68

Activations Density 0.000%

No Known Activations

This feature has no known activations.