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

agate

-0.75

atism

-0.74

nces

-0.70

ograms

-0.70

ogram

-0.68

isy

-0.67

section

-0.66

ias

-0.65

TTL

-0.65

orah

-0.65

POSITIVE LOGITS

 Flavoring

0.76

SpaceEngineers

0.70

EY

0.69

 Keen

0.68

 Pens

0.65

 Ships

0.64

EStreamFrame

0.63

 Core

0.62

 Rivals

0.62

 Races

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.