INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

heed

-0.80

awks

-0.79

ovember

-0.78

ager

-0.74

aving

-0.72

addons

-0.71

hid

-0.70

arcity

-0.70

acqu

-0.68

avery

-0.67

POSITIVE LOGITS

 Prosecut

0.75

rition

0.67

Collins

0.67

 Seraph

0.67

RIPT

0.64

endix

0.63

 Testament

0.63

vill

0.63

 Piper

0.62

Liter

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.