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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

enium

-0.74

rice

-0.72

usc

-0.65

uzz

-0.65

rim

-0.65

rier

-0.64

debian

-0.62

bleacher

-0.60

hib

-0.59

utton

-0.58

POSITIVE LOGITS

cause

0.81

 Cause

0.81

âĸ¬

0.73

ARP

0.72

PASS

0.71

Cause

0.68

acebook

0.67

APH

0.67

AME

0.65

binding

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.