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

 millenn

-0.88

aspers

-0.85

ĪĴ

-0.83

 rall

-0.83

unker

-0.81

 rave

-0.81

rower

-0.79

ibaba

-0.79

inventoryQuantity

-0.76

»Ĵ

-0.75

POSITIVE LOGITS

 Donation

0.72

 Achievement

0.70

 Amount

0.67

except

0.67

Commercial

0.65

 Fees

0.63

 Judicial

0.63

Exception

0.63

cases

0.62

 Vehicle

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.