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

 Blueprint

-0.72

 Commando

-0.68

 Agenda

-0.68

 Fant

-0.64

 Unreal

-0.64

 Puppet

-0.64

 Graves

-0.63

 Robot

-0.63

 Exhibit

-0.62

 Creative

-0.62

POSITIVE LOGITS

theless

0.94

soever

0.85

hemy

0.83

lish

0.79

hran

0.78

ccording

0.77

peed

0.75

rition

0.75

 behav

0.72

hered

0.71

Activations Density 0.000%

No Known Activations

This feature has no known activations.