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

lev

-0.80

-0.67

Ru

-0.66

LINE

-0.65

 shortage

-0.64

Upgrade

-0.64

accessible

-0.63

 entitlement

-0.62

TEXTURE

-0.61

igible

-0.61

POSITIVE LOGITS

ĸļ

0.88

 Parables

0.77

 psychologists

0.72

 Indonesia

0.70

 Disk

0.67

 Daly

0.67

plets

0.66

nep

0.66

 eleph

0.66

 Kuala

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.