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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

furt

-0.75

artment

-0.64

ioned

-0.63

founded

-0.62

ichick

-0.62

bery

-0.61

 grounding

-0.60

aunt

-0.59

-0.58

 Dempsey

-0.57

POSITIVE LOGITS

©¶æ¥µ

0.86

Downloadha

0.74

Untitled

0.73

pora

0.73

arios

0.70

Ĥª

0.68

xus

0.68

DragonMagazine

0.67

desktop

0.66

appa

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.