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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 crypt

-0.69

EStream

-0.68

 rotor

-0.67

 âĪ¼

-0.63

aer

-0.63

 cryptographic

-0.62

 Transparency

-0.61

COP

-0.61

 Burr

-0.61

 rall

-0.61

POSITIVE LOGITS

eta

0.86

orah

0.85

chio

0.82

amy

0.79

opol

0.78

zees

0.77

 Flavoring

0.76

anza

0.76

ita

0.76

stocks

0.75

Activations Density 0.000%

No Known Activations

This feature has no known activations.