INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

Reviewer

-0.78

DragonMagazine

-0.74

Course

-0.73

Commerce

-0.71

etimes

-0.70

 Skydragon

-0.69

netflix

-0.67

 comprom

-0.67

FactoryReloaded

-0.66

natureconservancy

-0.66

POSITIVE LOGITS

aed

0.69

ilk

0.68

oqu

0.67

Nah

0.62

 Jama

0.62

rir

0.61

 pseud

0.61

 Beir

0.59

Ack

0.58

dL

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.