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

ãħĭ

-0.77

 kcal

-0.76

+.

-0.71

 Probe

-0.69

ãħĭãħĭ

-0.68

 Venezuel

-0.68

 -----

-0.67

 humans

-0.66

 PEOPLE

-0.65

 Pengu

-0.63

POSITIVE LOGITS

inth

0.85

izontal

0.82

iHUD

0.70

iding

0.69

ŀ

0.66

 workaround

0.65

LEY

0.65

utory

0.64

native

0.63

 mesh

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.