INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

nexus

-0.67

vantage

-0.64

 bluff

-0.64

udi

-0.64

mol

-0.64

adian

-0.63

umenthal

-0.63

 shepherd

-0.62

 circulation

-0.61

rine

-0.61

POSITIVE LOGITS

Pon

0.73

Beg

0.73

 Dmit

0.72

 Canaver

0.71

Andre

0.70

Bey

0.70

 McKin

0.69

 Airl

0.68

 Krish

0.68

Grant

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.