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

 Plans

-0.80

Press

-0.75

 Press

-0.74

iod

-0.72

 Expect

-0.72

letes

-0.70

ROM

-0.69

Rush

-0.69

 Medals

-0.68

OME

-0.68

POSITIVE LOGITS

 anecd

0.73

 cloak

0.68

etooth

0.68

 distingu

0.62

 empir

0.62

arial

0.62

clo

0.62

tru

0.62

 sneaking

0.61

otomy

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.