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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ifax

-0.74

nda

-0.69

 surgeries

-0.66

LR

-0.66

 experien

-0.65

elvet

-0.63

 Walton

-0.61

uka

-0.61

atche

-0.61

ooth

-0.61

POSITIVE LOGITS

DragonMagazine

0.81

é»Ĵ

0.73

nat

0.70

$$

0.69

lessness

0.69

LECT

0.68

OD

0.67

 Incarnation

0.66

pread

0.64

FORE

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.