INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

Redditor

-0.68

paio

-0.68

pour

-0.65

entious

-0.65

cffffcc

-0.64

nir

-0.63

quickShipAvailable

-0.63

tackle

-0.62

ful

-0.62

ascript

-0.62

POSITIVE LOGITS

 Kennedy

0.69

 playbook

0.66

etus

0.63

 pedest

0.62

HHS

0.60

Uri

0.60

 Jacobs

0.59

 Destination

0.59

andum

0.58

 anatomy

0.57

Activations Density 0.000%

No Known Activations

This feature has no known activations.