INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 Maze

-0.71

ãĥĨ

-0.68

Cir

-0.67

NAT

-0.66

 Stall

-0.66

 Pepe

-0.62

stall

-0.60

itia

-0.60

igen

-0.60

 Juven

-0.59

POSITIVE LOGITS

elaide

0.72

anders

0.68

ictions

0.66

assian

0.65

ourn

0.64

orage

0.64

ear

0.62

Downloadha

0.62

eger

0.62

iction

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.