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

 srfAttach

-0.79

 Cantor

-0.75

usat

-0.74

IVERS

-0.71

 ---------

-0.71

alus

-0.69

 Heard

-0.68

uler

-0.65

icka

-0.64

 sacrific

-0.63

POSITIVE LOGITS

EMP

0.66

othing

0.63

],[

0.62

 poppy

0.62

apsed

0.60

opher

0.59

 Chapters

0.59

ienne

0.59

olate

0.58

 interrupt

0.57

Activations Density 0.000%

No Known Activations

This feature has no known activations.