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

OHN

-0.86

à¥

-0.78

olon

-0.69

à¤¾

-0.68

avascript

-0.65

Ô

-0.65

èĢħ

-0.62

olson

-0.61

quickShipAvailable

-0.61

ãĥĥ

-0.60

POSITIVE LOGITS

cise

0.71

 Helpful

0.67

 contrace

0.65

 WARN

0.64

rol

0.63

 disse

0.63

ailability

0.61

nice

0.61

lav

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.