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

 Kessler

-0.83

stro

-0.64

ngth

-0.63

minist

-0.62

USE

-0.61

 MÃ¼

-0.61

 Knot

-0.61

apo

-0.60

ignty

-0.58

 Gutenberg

-0.58

POSITIVE LOGITS

bestos

0.71

arus

0.71

Interstitial

0.69

agara

0.67

ghazi

0.67

Syrian

0.66

Ø¨

0.65

Audio

0.64

Baby

0.64

apest

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.