INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

onse

-0.68

astical

-0.68

 Dominican

-0.68

soDeliveryDate

-0.67

CCP

-0.66

ude

-0.66

aurus

-0.63

 Hindu

-0.61

 Buddha

-0.60

 princip

-0.60

POSITIVE LOGITS

hyde

0.72

 charger

0.70

Hunt

0.69

agnar

0.66

drivers

0.66

ufact

0.65

BOOK

0.64

"],"

0.64

payers

0.63

ware

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.