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

gage

-0.69

liga

-0.69

lied

-0.69

score

-0.68

ston

-0.66

 Belt

-0.64

 bidding

-0.64

fur

-0.63

knife

-0.63

furt

-0.63

POSITIVE LOGITS

aryn

0.74

çīĪ

0.69

oyal

0.67

ushima

0.67

 Both

0.65

abol

0.64

amph

0.63

 both

0.63

ull

0.63

usting

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.