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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

doms

-0.90

quickShipAvailable

-0.72

 Curve

-0.72

Winged

-0.71

emaker

-0.70

ograph

-0.68

esville

-0.67

rawdownloadcloneembedreportprint

-0.67

ypes

-0.65

TPPStreamerBot

-0.64

POSITIVE LOGITS

uld

0.87

 Gohan

0.69

inoa

0.64

 rehe

0.64

 Kling

0.63

Tai

0.62

 Clover

0.62

 Kaine

0.61

Pai

0.61

ãħĭãħĭ

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.