INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

ĺħ

-0.78

¿½

-0.77

waukee

-0.73

relation

-0.73

gio

-0.68

cled

-0.68

roxy

-0.67

arranted

-0.67

 Penalty

-0.67

Flor

-0.67

POSITIVE LOGITS

ï¸

0.70

 pens

0.69

elight

0.66

 whales

0.64

VPN

0.63

ebook

0.63

 Winc

0.62

asio

0.62

 trough

0.62

rooms

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.