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

ufact

-0.80

vernment

-0.75

 requested

-0.70

eks

-0.69

ffic

-0.68

phrine

-0.68

 redevelop

-0.68

avement

-0.68

 sidx

-0.67

Â·Â·

-0.67

POSITIVE LOGITS

 Paleo

0.81

Orig

0.74

bush

0.72

 Pascal

0.71

 Orion

0.69

mere

0.68

 âĵĺ

0.67

 Neal

0.67

 Nebula

0.67

aph

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.