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

puters

-0.70

 Commando

-0.68

 Lenovo

-0.66

HMS

-0.65

 SATA

-0.65

 Vlad

-0.64

 Kraken

-0.63

thinkable

-0.60

 MacBook

-0.60

DDR

-0.59

POSITIVE LOGITS

onson

0.73

olph

0.68

orial

0.67

coe

0.65

uds

0.65

speech

0.63

amiya

0.62

mun

0.61

ById

0.61

Ã©t

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.