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

zeb

-0.72

MET

-0.72

shows

-0.71

 carriers

-0.70

irmation

-0.68

 Holder

-0.67

quished

-0.67

 Carrier

-0.65

 Brush

-0.64

headers

-0.64

POSITIVE LOGITS

ngth

0.77

omething

0.70

yssey

0.69

jah

0.66

alg

0.64

 è£ıè¦ļéĨĴ

0.61

CVE

0.60

 disapp

0.59

 dece

0.59

 insult

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.