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

itude

-0.79

"]=>

-0.72

itudes

-0.68

":"","

-0.67

etry

-0.67

 Universe

-0.65

etric

-0.64

olan

-0.63

ito

-0.63

API

-0.63

POSITIVE LOGITS

 figured

0.67

 toured

0.66

 traveled

0.61

 premiered

0.61

 Dull

0.59

 preacher

0.59

accompan

0.59

 pastors

0.59

 cleric

0.58

 visas

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.