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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Garrison

-0.78

 sidx

-0.74

Op

-0.73

ahu

-0.69

OURCE

-0.68

 anew

-0.67

atan

-0.66

soDeliveryDate

-0.65

Pic

-0.64

 sqor

-0.64

POSITIVE LOGITS

currency

0.65

 buses

0.65

ocity

0.63

 recy

0.62

 governs

0.61

lessly

0.60

 travellers

0.60

 stewards

0.60

cycl

0.60

 cleaners

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.