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

bons

-0.77

ernels

-0.77

=>

-0.74

forcement

-0.73

nuts

-0.73

emate

-0.71

leys

-0.71

adj

-0.71

olate

-0.70

iband

-0.69

POSITIVE LOGITS

 Souls

0.69

 Genius

0.63

 illiter

0.62

Jub

0.62

 reper

0.61

Pru

0.60

 gracious

0.60

 curs

0.60

 Hitman

0.59

 Leviathan

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.