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

acus

-0.82

Īè

-0.80

 nails

-0.74

icum

-0.72

OAD

-0.69

atar

-0.68

heny

-0.67

OIL

-0.67

ndum

-0.66

YL

-0.65

POSITIVE LOGITS

 Discord

0.74

 Twisted

0.72

 Shack

0.67

wagon

0.66

 Carnival

0.63

 Valiant

0.62

 Haunted

0.61

 Robot

0.61

 Archdemon

0.61

 Blink

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.