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

imeters

-0.90

ullivan

-0.75

imeter

-0.75

Mos

-0.74

vernment

-0.72

ights

-0.71

 Schne

-0.70

ometimes

-0.70

pread

-0.66

cknowled

-0.66

POSITIVE LOGITS

PLA

0.77

ALE

0.72

 Odyssey

0.68

cx

0.67

 Grain

0.62

 Railroad

0.61

 Strikes

0.60

 HERO

0.60

crime

0.60

 Graphics

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.