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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ARI

-0.78

ante

-0.76

anie

-0.74

OTOS

-0.71

ographs

-0.68

KA

-0.68

Published

-0.64

utan

-0.63

éļ

-0.63

RON

-0.62

POSITIVE LOGITS

escape

0.72

agers

0.70

xit

0.70

 Tillerson

0.66

Py

0.64

alty

0.64

edia

0.63

selves

0.63

ãĤ®

0.63

riott

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.