INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Cel

-0.74

olars

-0.72

PHI

-0.69

 Indra

-0.69

 Fernand

-0.68

SER

-0.66

 Colleg

-0.66

 Crusher

-0.66

 Kinnikuman

-0.65

 discipl

-0.65

POSITIVE LOGITS

iHUD

0.73

tons

0.70

FTWARE

0.70

shows

0.69

ullivan

0.67

haven

0.66

 commits

0.66

asty

0.65

tube

0.64

hou

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.