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

licted

-0.74

ternity

-0.67

ificent

-0.64

cci

-0.64

ndra

-0.63

aundering

-0.62

 exoner

-0.61

ÅĤ

-0.61

sts

-0.60

 laund

-0.60

POSITIVE LOGITS

iculty

0.77

ience

0.66

DragonMagazine

0.65

ï¸

0.64

hip

0.61

pires

0.61

 Bark

0.59

yon

0.57

 Dudley

0.57

stakes

0.57

Activations Density 0.000%

No Known Activations

This feature has no known activations.