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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

é¾įå¥ĳå£«

-0.79

_>

-0.68

 Frieza

-0.66

ÏĢ

-0.65

 Pricing

-0.64

EStream

-0.64

 centimeters

-0.62

 Meridian

-0.62

 Kappa

-0.62

 Neptune

-0.61

POSITIVE LOGITS

ables

0.74

aks

0.68

able

0.67

ensable

0.67

ogh

0.66

 honoured

0.65

holders

0.63

gements

0.62

orsche

0.62

odge

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.