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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Christy

-0.68

 flares

-0.64

mc

-0.64

 Stellar

-0.62

 Tucson

-0.61

ahime

-0.61

opsis

-0.60

 McDonnell

-0.59

UCS

-0.59

 Kinn

-0.59

POSITIVE LOGITS

*/(

0.74

tail

0.72

 bona

0.67

lic

0.67

Gender

0.66

ulic

0.66

nil

0.66

uit

0.64

pter

0.63

itive

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.