INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

icion

-0.86

enegger

-0.76

dyl

-0.73

âĶ

-0.70

icht

-0.69

agne

-0.68

adian

-0.67

ensor

-0.66

*/(

-0.65

 answ

-0.64

POSITIVE LOGITS

 Christy

0.65

odi

0.63

 coron

0.63

 Generations

0.62

 Leilan

0.61

umeric

0.61

İĭ

0.60

MLA

0.60

ivation

0.59

 Pione

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.