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

yg

-0.75

resses

-0.70

orters

-0.68

geries

-0.66

izens

-0.63

marks

-0.63

ning

-0.62

ren

-0.62

ths

-0.62

anty

-0.61

POSITIVE LOGITS

ahime

0.77

wcs

0.76

amaz

0.74

Originally

0.65

DAQ

0.64

alf

0.64

 conduc

0.63

ECA

0.63

ava

0.61

 redes

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.