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

idge

-0.76

tti

-0.75

toe

-0.70

Grey

-0.68

wra

-0.67

hair

-0.65

skin

-0.64

iol

-0.64

yt

-0.63

irm

-0.63

POSITIVE LOGITS

itialized

0.83

DAQ

0.75

arnaev

0.74

soDeliveryDate

0.71

 Travels

0.71

æ©

0.68

 Skydragon

0.67

————————

0.66

 smartest

0.65

irez

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.