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

 Niet

-0.76

 Palest

-0.72

 princ

-0.68

 Columb

-0.67

rene

-0.66

 affili

-0.65

rams

-0.64

 Nationwide

-0.63

 Staten

-0.63

 Cornwall

-0.63

POSITIVE LOGITS

insky

0.77

 Mahjong

0.70

ho

0.68

 Potion

0.63

yp

0.63

yo

0.63

 Werewolf

0.63

ornings

0.63

 Explain

0.62

bat

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.