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

atever

-0.84

acus

-0.81

ascript

-0.72

 glor

-0.71

 dependence

-0.69

 facet

-0.68

ajo

-0.65

heid

-0.64

 compressor

-0.62

 coefficient

-0.61

POSITIVE LOGITS

ãĥ¡

0.85

rand

0.76

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

0.75

 Bringing

0.73

nces

0.71

 GNOME

0.70

nard

0.69

ãĥī

0.68

Ruk

0.67

assetsadobe

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.