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

 disadvantages

-0.76

ãģĻ

-0.66

 doub

-0.66

xit

-0.65

 distractions

-0.63

 demoral

-0.63

gey

-0.63

 humiliation

-0.61

Um

-0.61

 ration

-0.60

POSITIVE LOGITS

inav

0.74

bey

0.71

ertodd

0.65

 Pione

0.65

smanship

0.65

 AUTH

0.65

eco

0.65

achev

0.64

catentry

0.64

arius

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.