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

roman

-0.83

ourgeois

-0.79

senal

-0.76

 defensively

-0.67

 rife

-0.66

 abound

-0.65

arc

-0.65

antage

-0.64

 =================

-0.63

fusc

-0.63

POSITIVE LOGITS

atri

0.81

 Truman

0.66

Tel

0.66

 Siri

0.62

IRO

0.62

 husbands

0.62

 sang

0.61

arta

0.61

IPP

0.60

 Independence

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.