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

ricks

-0.74

raper

-0.72

aign

-0.71

Vec

-0.70

rez

-0.67

bley

-0.66

href

-0.64

pace

-0.64

eworks

-0.63

Loading

-0.62

POSITIVE LOGITS

ANC

0.79

oxin

0.77

ortunate

0.68

 Handling

0.67

lain

0.67

asio

0.66

IDES

0.65

estern

0.64

ctica

0.64

iosis

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.