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

 conclud

-0.79

 Appears

-0.72

 Notting

-0.71

âĸº

-0.71

enegger

-0.71

SPONSORED

-0.70

 quarters

-0.67

.ï¿½

-0.64

 grips

-0.63

 disarm

-0.63

POSITIVE LOGITS

heet

0.81

²¾

0.77

£

0.76

Ģ

0.71

ht

0.70

Ĥ¬

0.69

amin

0.69

uer

0.68

undown

0.68

į

0.68

Activations Density 0.000%

No Known Activations

This feature has no known activations.