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

monkey

-0.81

gaard

-0.79

-0.77

pret

-0.69

å§«

-0.69

amon

-0.69

cgi

-0.68

æ©

-0.68

ieri

-0.67

vice

-0.66

POSITIVE LOGITS

efeated

0.71

 Cortana

0.65

 Adren

0.64

OU

0.62

 streak

0.61

iphany

0.61

 revelation

0.60

 Trooper

0.60

Az

0.60

uala

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.