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

Ð¼

-0.73

eries

-0.71

ERY

-0.70

 Imperium

-0.70

Ó

-0.68

ilon

-0.68

clerosis

-0.67

hip

-0.66

Ļ

-0.65

dash

-0.64

POSITIVE LOGITS

 satur

0.72

que

0.69

 contrace

0.68

enture

0.68

asonable

0.67

 conclud

0.64

ques

0.63

 conscience

0.62

Uni

0.62

 contempor

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.