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

\\\\\\\\

-0.79

idation

-0.70

Ģ

-0.67

ocalypse

-0.67

ource

-0.67

roma

-0.67

Ci

-0.67

tainment

-0.66

Ĭ

-0.66

uthor

-0.66

POSITIVE LOGITS

bound

0.67

WM

0.66

 gazing

0.64

 Photographer

0.62

 ARTICLE

0.62

0.60

Spa

0.60

DW

0.59

awar

0.59

MpServer

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.