INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ãĤ´

-0.74

ãĥ³ãĤ¸

-0.63

 incent

-0.63

ãĤ«

-0.61

Ô

-0.61

Æ

-0.61

irin

-0.60

DES

-0.59

enh

-0.58

destruct

-0.57

POSITIVE LOGITS

ittens

0.71

ItemTracker

0.68

 Crimean

0.64

Nib

0.63

board

0.63

ibaba

0.62

 bureaucracy

0.61

habi

0.59

 largeDownload

0.59

atican

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.