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

ampions

-0.78

inic

-0.76

iage

-0.71

reciation

-0.67

saf

-0.65

igers

-0.64

 ceremony

-0.63

enic

-0.63

otent

-0.62

stros

-0.61

POSITIVE LOGITS

Snap

0.71

Connector

0.69

 Snapdragon

0.67

lake

0.67

ç¥ŀ

0.66

é¾įå

0.65

bats

0.65

bender

0.64

 Potato

0.64

terson

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.