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

NESS

-0.72

 ãĥ

-0.72

aps

-0.71

 Dominican

-0.71

æ

-0.66

jud

-0.64

hare

-0.63

auth

-0.63

apsed

-0.62

eded

-0.61

POSITIVE LOGITS

ulously

0.77

rily

0.77

strom

0.72

HL

0.70

emetery

0.67

terday

0.67

 stack

0.66

lar

0.66

ModLoader

0.65

anwhile

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.