INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

zzi

-0.78

aurus

-0.76

zinski

-0.71

ebra

-0.69

¥µ

-0.68

XT

-0.68

pez

-0.68

_-_

-0.67

Unknown

-0.66

chedel

-0.65

POSITIVE LOGITS

inventoryQuantity

0.85

manac

0.74

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

0.71

peer

0.66

 rating

0.65

lust

0.64

 thinly

0.63

 ponder

0.63

loo

0.62

 mandate

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.