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

ashtra

-0.65

VW

-0.64

 jealous

-0.62

WN

-0.61

 ejac

-0.61

 intoxication

-0.60

 arthritis

-0.58

monkey

-0.57

HUD

-0.57

 schizophrenia

-0.56

POSITIVE LOGITS

 Layer

0.73

 Orchestra

0.72

 Inqu

0.70

 Writ

0.68

ieth

0.67

 Faster

0.66

 Vigil

0.65

 Conce

0.65

 Pref

0.64

 pamph

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.