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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

psc

-0.86

spir

-0.75

arrow

-0.71

Flag

-0.71

icated

-0.70

bucks

-0.69

dose

-0.69

inventoryQuantity

-0.68

illac

-0.68

iotics

-0.68

POSITIVE LOGITS

 Nicole

0.72

 Papua

0.72

 Bakr

0.67

Phi

0.67

 Sloven

0.67

 Maria

0.65

 Julio

0.64

 Julius

0.64

 Dirk

0.63

 Slovenia

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.