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

xxxxxxxx

-0.73

atus

-0.71

yll

-0.70

-0.69

VP

-0.68

alse

-0.67

las

-0.65

(-

-0.64

JP

-0.64

('

-0.64

POSITIVE LOGITS

 carbohyd

0.64

æĸ¹

0.64

 imitation

0.64

 treadmill

0.63

 differential

0.61

 Wake

0.60

finger

0.60

raint

0.59

 amplification

0.58

 Wheeler

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.