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

Lug

-0.78

 Â®

-0.77

anu

-0.68

 Singapore

-0.68

cept

-0.63

 Jama

-0.62

ARDIS

-0.61

Taj

-0.61

 Luxembourg

-0.61

 Mald

-0.61

POSITIVE LOGITS

 predec

0.76

natureconservancy

0.73

©¶æ

0.72

rentices

0.72

 utilizes

0.71

luaj

0.71

aturdays

0.71

ĳå£«

0.66

igent

0.66

ç«

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.