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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Kubrick

-0.78

unity

-0.70

uda

-0.67

mania

-0.66

ha

-0.65

alf

-0.65

hub

-0.64

windows

-0.64

 Scha

-0.64

NBC

-0.63

POSITIVE LOGITS

iott

0.75

 sidx

0.70

 cryptoc

0.68

è¦ļéĨĴ

0.67

cano

0.64

abama

0.62

nesota

0.60

 sqor

0.59

oller

0.59

eous

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.