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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

quished

-0.84

 jams

-0.82

Ļ½

-0.78

 Travels

-0.77

«ĺ

-0.75

¶

-0.74

aneers

-0.73

¶ħ

-0.72

Shot

-0.71

²¾

-0.70

POSITIVE LOGITS

isen

0.72

resa

0.70

ero

0.69

hing

0.67

Pok

0.66

igue

0.66

olerance

0.64

finder

0.64

agn

0.62

 suit

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.