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

Die

-0.70

Fal

-0.69

Craft

-0.67

aughs

-0.67

 Sack

-0.66

 disinfect

-0.64

Guard

-0.64

Gal

-0.64

 Cosmetic

-0.63

 respir

-0.63

POSITIVE LOGITS

tem

0.72

 repayment

0.71

»Ĵ

0.70

»

0.65

 outsourcing

0.63

 citiz

0.62

oi

0.62

 remod

0.62

 ballpark

0.61

 Miko

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.