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

 Opportun

-0.69

aughs

-0.66

Labour

-0.64

 Spit

-0.63

POV

-0.63

keye

-0.62

 Weir

-0.62

Nat

-0.61

tek

-0.61

 Sussex

-0.60

POSITIVE LOGITS

quez

0.70

azo

0.68

uchi

0.67

soDeliveryDate

0.66

oul

0.66

enthal

0.64

 clipping

0.64

icals

0.64

ãĤĬ

0.62

chu

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.