INDEX

Explanations

words associated with gender, style, and fashion

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

År

-0.86

 saar

-0.85

 Sén

-0.84

 franz

-0.83

 gallina

-0.83

 kram

-0.83

 meis

-0.83

 pank

-0.82

 incess

-0.81

 palme

-0.81

POSITIVE LOGITS

AndEndTag

0.58

XtraReports

0.56

ruly

0.54

relenting

0.54

 autorytatywna

0.53

mistak

0.52

 pinulongan

0.52

thinkable

0.52

المناصب

0.51

sightly

0.51

Activations Density 0.531%

words associated with gender, style, and fashion

No Comments

No Known Activations