INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

ItemTracker

-0.75

 Territories

-0.66

çĶ

-0.64

Benz

-0.62

REDACTED

-0.62

Charges

-0.61

å¤

-0.61

SAM

-0.60

Wisconsin

-0.59

 rights

-0.59

POSITIVE LOGITS

nect

0.83

clave

0.79

jack

0.71

olate

0.70

fit

0.69

ftime

0.66

iphate

0.65

 populate

0.65

icate

0.65

enei

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.