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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

REE

-0.89

AH

-0.68

ail

-0.67

irm

-0.66

heses

-0.65

che

-0.65

PSU

-0.64

afety

-0.64

wd

-0.64

Awesome

-0.63

POSITIVE LOGITS

taboola

0.66

 Circus

0.60

 msec

0.59

 Dunk

0.58

 occupies

0.57

âĶĢâĶĢâĶĢâĶĢ

0.57

soType

0.56

Ï

0.55

 tabletop

0.55

 Invaders

0.55

Activations Density 0.000%

No Known Activations

This feature has no known activations.