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

bol

-0.85

 Pengu

-0.75

rador

-0.73

Downloadha

-0.72

atche

-0.72

 guiActiveUn

-0.71

Educ

-0.67

valued

-0.67

yrinth

-0.67

 TAMADRA

-0.67

POSITIVE LOGITS

ellow

0.77

 goodbye

0.76

 Breakfast

0.61

 corpor

0.59

leck

0.59

 report

0.58

 breakfast

0.58

 Wesley

0.57

aths

0.57

 Spielberg

0.56

Activations Density 0.000%

No Known Activations

This feature has no known activations.