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

thro

-0.88

tis

-0.78

 Slave

-0.77

 Shades

-0.74

slave

-0.72

seed

-0.69

til

-0.67

 Seek

-0.65

 Venezuel

-0.63

Hed

-0.63

POSITIVE LOGITS

otos

0.78

ocally

0.78

 Koen

0.72

Downloadha

0.71

ģ«

0.69

VIDIA

0.68

ike

0.68

VERTISEMENT

0.65

oker

0.64

ettings

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.