INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 ethic

-0.68

 showc

-0.67

iets

-0.66

 stamps

-0.66

 desk

-0.66

 glue

-0.65

ife

-0.64

nice

-0.62

 Line

-0.62

interface

-0.60

POSITIVE LOGITS

 Franch

1.08

ruary

0.89

alam

0.77

 Luxem

0.75

Kad

0.74

auga

0.73

olin

0.68

theless

0.66

imately

0.65

 Ferr

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.