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

©¶æ

-0.94

 destro

-0.87

å°Ĩ

-0.87

 millenn

-0.79

 refrain

-0.79

ntil

-0.78

 rall

-0.77

 occas

-0.76

 istg

-0.74

ĸļ

-0.70

POSITIVE LOGITS

naire

0.66

washing

0.64

PNG

0.64

 expands

0.63

Moh

0.62

Fal

0.62

 Military

0.60

 Plains

0.60

wood

0.60

ouls

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.