INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

unes

-0.80

ights

-0.77

src

-0.76

ungle

-0.72

Raw

-0.69

Interstitial

-0.69

pur

-0.68

ub

-0.68

ayn

-0.65

inventoryQuantity

-0.65

POSITIVE LOGITS

ensical

0.73

 paternity

0.72

 reversible

0.71

 forgiveness

0.65

phony

0.65

 fuss

0.64

isma

0.64

 Twist

0.63

 Sloan

0.62

cision

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.