INDEX

Explanations

numbers and special characters used in phrases

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Configuration

jbloom/Gemma-2b-IT-Residual-Stream-SAEs/gemma_2b_it_blocks.12.hook_resid_post_16384

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

HuggingFaceFW/fineweb

Features

16,384

Data Type

float32

Hook Name

blocks.12.hook_resid_post

Hook Layer

Architecture

standard

Context Size

1,024

Dataset

Skylion007/openwebtext

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 kram

-1.74

 utop

-1.70

 glan

-1.55

 pessi

-1.54

 abnorm

-1.53

hek

-1.50

 bayern

-1.49

 elek

-1.48

 makro

-1.47

zyn

-1.46

POSITIVE LOGITS

 Apesar

0.82

Özel

0.78

Çok

0.76

Açık

0.74

Ayrıca

0.74

yaitu

0.70

 Mesmo

0.67

FlatAppearance

0.67

Selanjutnya

0.67

 Após

0.67

Activations Density 0.626%

numbers and special characters used in phrases

No Comments

No Known Activations