INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

rogens

-0.78

 contrace

-0.78

pas

-0.77

 paran

-0.73

awaru

-0.73

netflix

-0.69

Abstract

-0.68

activ

-0.67

 guiActiveUn

-0.66

Austral

-0.66

POSITIVE LOGITS

 Breaker

0.68

 Rebels

0.66

 depth

0.65

 Memor

0.60

 wings

0.59

 Cance

0.59

 lane

0.58

 Bread

0.58

 Codec

0.58

Typ

0.57

Activations Density 0.000%

No Known Activations

This feature has no known activations.