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

recy

-0.74

quartered

-0.68

 subclass

-0.68

Tags

-0.66

core

-0.65

ettings

-0.63

Shell

-0.62

Understanding

-0.62

wik

-0.62

clud

-0.62

POSITIVE LOGITS

IFE

0.73

ITIES

0.72

ITNESS

0.71

llah

0.69

ESV

0.68

 Elise

0.67

quickShipAvailable

0.66

 Tribal

0.66

BILITY

0.64

ITED

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.