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

CPS

-0.72

 Trayvon

-0.66

 Lank

-0.63

 crore

-0.62

 Breach

-0.62

ãĥ¥

-0.61

cc

-0.60

 ï¿½ï¿½ï¿½ï¿½

-0.60

 Million

-0.60

+++

-0.59

POSITIVE LOGITS

emis

0.81

pmwiki

0.81

ebook

0.74

ubi

0.74

 regiment

0.72

ourses

0.70

yip

0.67

ryu

0.65

DragonMagazine

0.65

atche

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.