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

itionally

-0.80

£ı

-0.69

onew

-0.69

 ingred

-0.67

 datas

-0.66

soDeliveryDate

-0.65

readable

-0.65

 careful

-0.63

©¶æ

-0.63

 likeness

-0.62

POSITIVE LOGITS

 Fiction

0.72

conom

0.70

tul

0.69

antha

0.69

IDS

0.68

 Perez

0.67

sth

0.67

âĵĺ

0.66

odes

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.