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

IGHTS

-0.63

å¸

-0.62

rpm

-0.61

igan

-0.61

 Beir

-0.59

 licence

-0.58

RON

-0.58

 faire

-0.57

=-=-

-0.56

Ang

-0.56

POSITIVE LOGITS

atown

0.75

irlf

0.72

ydia

0.72

atches

0.72

earable

0.72

adj

0.69

 Cortana

0.68

aura

0.67

Seattle

0.66

buquerque

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.