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

 Technician

-0.77

Catholic

-0.68

 Scientist

-0.65

 Investor

-0.64

aurus

-0.63

 Letters

-0.63

 Patriot

-0.61

baugh

-0.60

 sear

-0.60

 Fighter

-0.60

POSITIVE LOGITS

mble

0.80

uras

0.78

estyles

0.75

ramid

0.73

 Kard

0.70

heid

0.69

 metab

0.69

ourses

0.68

 withd

0.68

 Seym

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.