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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 advoc

-0.81

ortium

-0.75

beard

-0.73

bj

-0.71

arij

-0.70

GBT

-0.68

 Catalyst

-0.67

isSpecialOrderable

-0.66

java

-0.66

DOM

-0.66

POSITIVE LOGITS

 ante

0.83

 squares

0.68

 dozen

0.63

 acre

0.63

pp

0.62

 Crimean

0.62

 noon

0.62

 batter

0.62

 Noon

0.62

 half

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.