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

Downloadha

-0.77

SSL

-0.77

Reviewer

-0.71

ãĤ¡

-0.70

WD

-0.70

 hiber

-0.69

 Kodi

-0.66

dating

-0.66

 messaging

-0.65

DVD

-0.64

POSITIVE LOGITS

 filled

0.85

 room

0.67

ulhu

0.67

 Sheet

0.67

anooga

0.67

 packed

0.66

estone

0.65

 Dealer

0.64

inous

0.63

indu

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.