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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

qt

-0.72

 Synt

-0.67

ospace

-0.66

okÃ©

-0.65

 workspace

-0.63

plates

-0.62

 drafts

-0.59

SF

-0.59

tem

-0.58

avia

-0.57

POSITIVE LOGITS

artney

0.87

TPPStreamerBot

0.75

earchers

0.73

aucas

0.70

ilater

0.69

Honest

0.69

Detailed

0.68

leigh

0.67

billion

0.67

patrick

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.