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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Simulator

-0.73

 Brigade

-0.72

 Vengeance

-0.68

ACTED

-0.67

 Turtle

-0.66

 Corsair

-0.66

 Lans

-0.65

 Weir

-0.64

Nou

-0.63

 Hare

-0.63

POSITIVE LOGITS

instein

0.82

gans

0.75

henko

0.73

 sovere

0.69

gypt

0.69

psey

0.68

onite

0.66

dist

0.65

uten

0.65

itar

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.