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

aturdays

-0.87

ï¸

-0.81

ISTER

-0.74

qt

-0.66

||||

-0.66

ideon

-0.65

annis

-0.65

reens

-0.64

istries

-0.64

Ã¼r

-0.63

POSITIVE LOGITS

arna

0.68

ATF

0.66

uci

0.65

 royalty

0.64

 reperc

0.63

RG

0.63

ilial

0.61

CRA

0.61

 Reneg

0.59

 shenan

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.