INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

riott

-0.81

ction

-0.73

hess

-0.73

âĸĪâĸĪâĸĪâĸĪ

-0.72

irc

-0.71

allo

-0.69

igue

-0.68

angible

-0.67

zona

-0.67

ochet

-0.66

POSITIVE LOGITS

 haul

0.69

 deadline

0.68

 trek

0.68

sag

0.66

 tales

0.65

 tits

0.64

 shotguns

0.63

 gull

0.63

 trips

0.62

 skinny

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.