INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

yss

-0.82

hement

-0.80

alty

-0.78

erie

-0.77

mine

-0.71

emen

-0.70

aline

-0.68

cients

-0.66

oran

-0.64

roid

-0.64

POSITIVE LOGITS

 Fedora

0.75

 Semin

0.73

 Casting

0.68

anan

0.67

 Doct

0.65

 Prometheus

0.65

 Percentage

0.63

 Steal

0.62

 Ment

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.