INDEX

Explanations

phone numbers and related 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

 elek

-1.41

 alkoh

-1.38

 maksi

-1.37

 panik

-1.36

uhr

-1.34

 saar

-1.34

 makro

-1.30

 antik

-1.29

 kram

-1.29

 kosme

-1.27

POSITIVE LOGITS

mädchen

0.69

 Ename

0.62

 Etching

0.59

Dlaczego

0.59

 kwiet

0.57

 impractica

0.57

 séjours

0.57

 nadzie

0.56

 Confidentiality

0.56

 Engraved

0.55

Activations Density 0.243%

phone numbers and related phrases

No Comments

No Known Activations