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

bringer

-0.76

sequence

-0.71

orate

-0.71

bench

-0.69

 Leban

-0.69

taker

-0.67

apeshifter

-0.67

ozyg

-0.67

iple

-0.66

lights

-0.63

POSITIVE LOGITS

 Wonderland

0.69

 PokÃ©

0.65

clus

0.63

TPP

0.63

 Adventures

0.63

 Legislation

0.62

ALE

0.62

apter

0.62

 Autism

0.61

aldo

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.