INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 lease

-0.70

 untreated

-0.68

ħĭ

-0.65

 Hull

-0.65

 panel

-0.64

rawdownloadcloneembedreportprint

-0.64

assic

-0.63

 skim

-0.62

 pads

-0.62

 homes

-0.61

POSITIVE LOGITS

Course

0.78

Winner

0.75

otle

0.75

ista

0.74

allo

0.74

Robin

0.73

reasonable

0.71

Ocean

0.70

things

0.69

PokÃ©mon

0.69

Activations Density 0.000%

No Known Activations

This feature has no known activations.