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

 holiest

-0.88

ighth

-0.85

 Palest

-0.85

abee

-0.80

eday

-0.78

ntil

-0.77

 toile

-0.75

edIn

-0.75

oday

-0.73

enment

-0.71

POSITIVE LOGITS

UCT

0.73

exp

0.71

FIN

0.67

BILITY

0.66

 broker

0.66

 clause

0.63

 Strikes

0.62

 Carbon

0.61

TER

0.61

APTER

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.