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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

realDonaldTrump

-0.79

Pad

-0.74

Rom

-0.69

 Quantity

-0.68

 Cosponsors

-0.67

 grounding

-0.64

DERR

-0.63

 ItemLevel

-0.63

unciation

-0.62

SourceFile

-0.62

POSITIVE LOGITS

afia

0.83

 territ

0.75

 Swordsman

0.72

erous

0.72

byter

0.71

 steals

0.70

igent

0.69

 Trojan

0.66

 compens

0.66

 eagle

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.