INDEX

Explanations

phrases related to specific technical information or model names

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

miyor

-0.85

iyor

-0.75

miyorum

-0.67

mişti

-0.63

artney

-0.61

diğimiz

-0.58

diğ

-0.56

Allister

-0.56

diğinde

-0.56

diğini

-0.55

POSITIVE LOGITS

 Mezz

0.97

 mezz

0.95

Me

0.95

Me

0.94

 ecru

0.91

 Megal

0.90

 compréhen

0.90

mef

0.88

Meg

0.87

MEG

0.86

Activations Density 0.172%

phrases related to specific technical information or model names

No Comments

No Known Activations