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

inventoryQuantity

-0.75

nown

-0.72

 mouths

-0.69

aldo

-0.65

 disadvantage

-0.65

ometown

-0.65

 simultane

-0.65

 disadvantages

-0.65

 publisher

-0.65

izens

-0.64

POSITIVE LOGITS

PH

0.70

 Marse

0.70

Å

0.65

OUS

0.64

Greek

0.64

Î½

0.64

REE

0.64

¸

0.63

chemical

0.62

GEN

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.