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

TERN

-0.71

 CLSID

-0.67

 availability

-0.65

MSN

-0.65

 Mund

-0.63

 Scotia

-0.63

 Sach

-0.63

Poc

-0.62

 Courtney

-0.61

 Tasman

-0.61

POSITIVE LOGITS

animous

0.86

interstitial

0.83

ards

0.73

achev

0.68

rouse

0.67

elected

0.65

Anonymous

0.64

gow

0.64

ingers

0.63

wire

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.