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

sein

-0.75

 Carbuncle

-0.73

 Gleaming

-0.71

 Frontier

-0.71

 Massacre

-0.70

 Realms

-0.70

 Unle

-0.69

Yan

-0.68

 Bloody

-0.67

 Jolly

-0.67

POSITIVE LOGITS

amera

0.77

odox

0.77

etheless

0.70

acea

0.69

 costly

0.67

HUD

0.64

acci

0.62

©¶æ

0.61

 captcha

0.60

outube

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.