INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

NetMessage

-0.97

nice

-0.76

ournal

-0.71

kind

-0.68

wikipedia

-0.66

nel

-0.65

Ó

-0.65

minster

-0.64

 Puppet

-0.63

Wikipedia

-0.63

POSITIVE LOGITS

 Calories

0.71

ŀ

0.67

Submit

0.62

ãĥ©ãĥ³

0.61

EVs

0.61

 bulls

0.60

imb

0.60

otaur

0.59

bot

0.59

 miscarriage

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.