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

ratulations

-0.88

rament

-0.81

bable

-0.80

itialized

-0.76

eds

-0.75

 âĶľ

-0.74

rious

-0.70

omsky

-0.70

 ACTIONS

-0.69

aroo

-0.68

POSITIVE LOGITS

BuyableInstoreAndOnline

0.78

 Faul

0.73

 Gutenberg

0.71

 Purg

0.70

 Aram

0.70

 Kafka

0.69

 Tolkien

0.69

 Herz

0.68

 Carib

0.67

 Lovecraft

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.