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

ibal

-0.86

ItemTracker

-0.84

rats

-0.84

heastern

-0.81

phans

-0.79

rals

-0.78

ao

-0.77

ahs

-0.76

 attm

-0.74

aye

-0.70

POSITIVE LOGITS

 caption

0.68

 DISTR

0.67

 variation

0.65

 denomin

0.64

 wavelength

0.64

 tint

0.63

PI

0.63

ISS

0.63

 deviation

0.62

DEN

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.