INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

portion

-0.80

owship

-0.73

peer

-0.73

omer

-0.72

port

-0.72

venge

-0.70

minist

-0.70

ges

-0.68

ordinate

-0.68

kin

-0.67

POSITIVE LOGITS

Ĥª

0.74

 easing

0.71

sonian

0.68

 adjourn

0.67

 anecd

0.66

jury

0.65

 expiration

0.64

İĭ

0.63

fib

0.62

 obstruction

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.