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

pain

-0.69

bound

-0.65

 heals

-0.65

 antibody

-0.65

nec

-0.62

 marrow

-0.62

DIV

-0.61

cember

-0.60

binding

-0.59

omy

-0.58

POSITIVE LOGITS

 Cheong

0.81

atche

0.75

Own

0.68

undown

0.67

untled

0.65

andals

0.65

ãĥķãĤ¡

0.65

 Cosponsors

0.64

DragonMagazine

0.64

ãĥķãĤ©

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.