Les modèles à raisonnement sont entraînés pour générer le "processus de réflexion" que le modèle suit pour générer sa réponse. Par conséquent, ils peuvent faire preuve de capacités de raisonnement supérieures dans leurs réponses par rapport aux modèles de base équivalents.
Le processus de réflexion est activé par défaut. Lorsque vous utilisez Vertex AI Studio, vous pouvez afficher l'intégralité du processus de réflexion, ainsi que la réponse générée par le modèle.
Modèles compatibles
Le processus de réflexion est compatible avec les modèles suivants :
Utiliser un modèle de réflexion
Pour utiliser la réflexion avec un modèle compatible :
Console
- Ouvrez Vertex AI Studio > Créer une requête.
- Dans le panneau Modèle, cliquez sur Changer de modèle, puis sélectionnez l'un des modèles compatibles dans le menu.
- (Gemini 2.5 Flash uniquement) Le budget de réflexion est défini sur Automatique par défaut lorsque le modèle est chargé.
- (Facultatif) Dans le champ Instructions système, donnez au modèle des instructions détaillées sur la mise en forme de ses réponses.
- Saisissez une requête dans le champ Écrivez votre requête.
- Cliquez sur Exécuter.
Gemini renvoie une réponse une fois qu'elle est générée. Selon la complexité de la réponse, ce processus peut prendre plusieurs secondes.
(Gemini 2.5 Flash uniquement) Pour désactiver la réflexion, définissez Budget de réflexion sur Désactivé.
Python
Installer
pip install --upgrade google-genai
Pour en savoir plus, lisez la documentation de référence du SDK.
Définissez les variables d'environnement pour utiliser le SDK Gen AI avec Vertex AI :
# Replace the `GOOGLE_CLOUD_PROJECT` and `GOOGLE_CLOUD_LOCATION` values # with appropriate values for your project. export GOOGLE_CLOUD_PROJECT=GOOGLE_CLOUD_PROJECT export GOOGLE_CLOUD_LOCATION=global export GOOGLE_GENAI_USE_VERTEXAI=True
Afficher les résumés des réflexions
Les résumés de réflexion sont la version abrégée du processus de réflexion que le modèle a suivi pour générer sa réponse. Vous pouvez afficher des résumés de pensées dans Gemini 2.5 Flash et Gemini 2.5 Pro. Pour afficher les résumés de réflexion, procédez comme suit :
Console
Les résumés de réflexion sont activés par défaut dans Vertex AI Studio. Vous pouvez voir le résumé du processus de réflexion du modèle en développant le panneau Étapes du raisonnement.
Python
Installer
pip install --upgrade google-genai
Pour en savoir plus, lisez la documentation de référence du SDK.
Définissez les variables d'environnement pour utiliser le SDK Gen AI avec Vertex AI :
# Replace the `GOOGLE_CLOUD_PROJECT` and `GOOGLE_CLOUD_LOCATION` values # with appropriate values for your project. export GOOGLE_CLOUD_PROJECT=GOOGLE_CLOUD_PROJECT export GOOGLE_CLOUD_LOCATION=global export GOOGLE_GENAI_USE_VERTEXAI=True
Contrôler le budget de réflexion
Vous pouvez contrôler la limite de réflexion du modèle lors de ses réponses. Cette limite supérieure est appelée budget de réflexion et s'applique à l'ensemble du processus de réflexion du modèle. Par défaut, le modèle contrôle automatiquement la quantité de jetons qu'il peut générer (jusqu'à un maximum de 8 192 jetons).
Vous pouvez définir manuellement la limite supérieure du nombre de jetons dans les situations où vous avez besoin de plus ou de moins de jetons que le budget de réflexion par défaut. Vous pouvez définir une limite de jetons plus faible pour les tâches moins complexes ou une limite plus élevée pour les tâches plus complexes.
Le tableau suivant indique les montants minimum et maximum que vous pouvez définir pour le budget de jetons pour chaque modèle compatible :
Modèle | Montant minimal de jetons | Montant maximal de jetons |
---|---|---|
Gemini 2.5 Flash | 1 | 24,576 |
Gemini 2.5 Pro | 128 | 32 768 |
Gemini 2.5 Flash-Lite | 512 | 24,576 |
Si vous définissez le budget de réflexion sur 0
lorsque vous utilisez Gemini 2.5 Flash et Gemini 2.5 Flash-Lite, la réflexion est désactivée. Il n'est pas possible de désactiver la réflexion pour Gemini 2.5 Pro.
Si vous souhaitez que le modèle contrôle le budget de réflexion lorsque vous utilisez l'API, définissez le budget de réflexion sur -1
.
Console
- Ouvrez Vertex AI Studio > Créer une requête.
- Dans le panneau Modèle, cliquez sur Changer de modèle, puis sélectionnez l'un des modèles compatibles dans le menu.
- Sélectionnez Manuel dans le sélecteur de menu déroulant Budget de réflexion, puis ajustez la limite à l'aide du curseur.
Python
Installer
pip install --upgrade google-genai
Pour en savoir plus, lisez la documentation de référence du SDK.
Définissez les variables d'environnement pour utiliser le SDK Gen AI avec Vertex AI :
# Replace the `GOOGLE_CLOUD_PROJECT` and `GOOGLE_CLOUD_LOCATION` values # with appropriate values for your project. export GOOGLE_CLOUD_PROJECT=GOOGLE_CLOUD_PROJECT export GOOGLE_CLOUD_LOCATION=global export GOOGLE_GENAI_USE_VERTEXAI=True
Techniques de rédaction de requêtes
Pour tirer le meilleur parti des modèles à raisonnement Gemini, il est essentiel de créer des requêtes efficaces. Cette section décrit quelques bonnes pratiques de rédaction de requêtes pour les modèles à raisonnement.
Commencez par des instructions générales décrivant le résultat souhaité. Commencez par une requête simple et observez la réflexion initiale du modèle sur la manière de répondre.
Describe the most unusual or striking architectural feature you've ever seen in a building or structure. <...> The most unusual architectural feature that has truly stayed with me is the "floating roof" of the Heydar Aliyev Center in Baku, Azerbaijan, designed by Zaha Hadid Architects. Visually, it's breathtaking. Imagine a vast, undulating landscape of pure white, seemingly draped over the building. It's not just a roof in the traditional sense; it's an extension of the ground, flowing seamlessly upwards and over the building's various spaces before curving back down to meet the earth again. There are no sharp angles, no rigid lines. Instead, everything is fluid, organic, and continuous. It resembles a giant, pristine white wave frozen in motion, or perhaps a sand dune sculpted by an unseen wind. <...> |
Fournir des instructions détaillées
Si la première tentative ne donne pas le résultat que vous souhaitez, affinez votre requête. Fournissez des instructions détaillées plus précises afin de guider Gemini dans le processus de résolution des problèmes pour une approche plus ciblée. Cela vous permet d'orienter la compréhension initiale de la tâche par le modèle avec des suggestions et des instructions précises.
Now imagine that architectural feature is not just for looks, but has a specific function or purpose related to the building's environment or the people who use it. Explain what that function could be and how it enhances the building's overall design and usability. <...> Let's imagine that the "floating roof" of the Heydar Aliyev Center is not purely aesthetic but serves a crucial environmental and user-centric function: it's designed as a sophisticated climate moderator and natural ventilation system for the Baku environment. Here's how this function would work and enhance the building: Function: Passive Climate Control and Natural Ventilation System Shading and Solar Heat Mitigation: The undulating curves of the white GFRC roof are strategically shaped and oriented to provide optimal shading throughout the day, year-round. Baku experiences intense sun, especially during the summer. The overhanging curves, acting like giant brise-soleils, would block direct sunlight from hitting the building's glass facade, significantly reducing solar heat gain. The varying heights and depths of the curves would create dynamic shade patterns, ensuring that different parts of the building are shaded at different times of the day, optimizing comfort and reducing the need for excessive air conditioning. The white color of the GFRC further enhances this by reflecting a large portion of the solar radiation. <...> |
Associer la rédaction de requêtes multi-shot à la réflexion
En associant la rédaction de requêtes multi-shot à la réflexion, vous pouvez améliorer la compréhension de Gemini et la précision de ses réponses. Fournissez plusieurs exemples de paires d'entrée-sortie, illustrant le comportement et le format de sortie souhaités dans votre requête.
Example 1: User: What is the tallest mountain in the world? Assistant: Mount Everest Example 2: User: What is the largest ocean? Assistant: Pacific Ocean User: What is the longest river in the world? Assistant: Amazon River |
Définir la sortie et le comportement
Lorsque vous créez une application dans laquelle les utilisateurs interagissent directement avec le modèle, vous pouvez fournir des conseils sur la façon dont la sortie et les réponses de Gemini seront formulées et formatées.
Instructions système
Les instructions système sont un ensemble d'instructions que le modèle traite avant de traiter les requêtes. Elles sont rappelées chaque fois qu'une requête est envoyée au modèle, et indiquent au modèle comment il doit se comporter et répondre. Par exemple, vous pouvez inclure des éléments tels que le rôle ou la persona, des informations contextuelles et des instructions de mise en forme.
System Instruction: You are a helpful and concise coding tutor. Always provide code examples and explain them briefly. User Prompt: Explain what is a closure in Javascript and give an example. function outerFunction(outerVar) { function innerFunction(innerVar) { console.log(outerVar); // Accessing outerVar from outerFunction's scope console.log(innerVar); } return innerFunction; } const closureExample = outerFunction('I am from outer'); closureExample('I am from inner'); Explanation: A closure is created when a function (innerFunction in this example) is defined inside another function (outerFunction) and the inner function remembers and has access to the variables from the outer function's scope, even after the outer function has finished executing. In the example: outerFunction is called with 'I am from outer' and returns innerFunction. closureExample now holds a reference to innerFunction. When closureExample('I am from inner') is called, innerFunction executes. Even though outerFunction has already finished, innerFunction can still access and use outerVar (which is 'I am from outer') because of the closure. In essence, the innerFunction "closes over" the variables of its outer function's scope, creating a closure. |
Vérification et réflexion
Incluez des étapes de validation spécifiques dans les instructions système ou directement dans vos requêtes. Par exemple, demandez à Gemini de vérifier ses sources, d'examiner son raisonnement, d'identifier les erreurs potentielles et de vérifier sa réponse finale.
Verify the following statement and provide your sources: The capital of Brazil is Rio de Janeiro. The statement "The capital of Brazil is Rio de Janeiro" is incorrect. The current capital of Brazil is Brasília. Rio de Janeiro was the capital of Brazil for a significant period, from 1763 to 1960. However, the capital was moved to Brasília in 1960 as part of a plan to develop the country's interior and create a more centrally located capital. Verification and Sources: You can easily verify this information from numerous reliable sources: Encyclopedias: Britannica: Searching for "Brazil" on Britannica will clearly state Brasília as the capital. https://www.britannica.com/place/Brazil (Look for the "Capital" information in the introductory summary or "Geography" section) Wikipedia: The Wikipedia article for "Brazil" also states Brasília as the capital. https://en.wikipedia.org/wiki/Brazil (Look in the infobox or introductory section) Government and Official Sources: CIA World Factbook: This is a highly reliable source for country information. Searching for "Brazil" will list Brasília as the capital. https://www.cia.gov/the-world-factbook/countries/brazil/ (Look under "Government" > "Capital") Official Website of the Brazilian Government (if accessible and in English): While navigating official government websites can sometimes be complex, they would definitively list Brasília as the capital. Geographic Resources: National Geographic: Their website and publications will confirm Brasília as the capital of Brazil. Atlases and Geography Textbooks: Any modern atlas or geography textbook will list Brasília as the capital. Why the confusion? The confusion might arise because: Rio de Janeiro is a very famous Brazilian city: It's a major tourist destination, known for its beaches, Carnival, and iconic landmarks like Christ the Redeemer. Its past status as capital adds to its historical significance and public recognition. Historical context: Rio de Janeiro was the capital for a long time. People who learned about Brazil in the past might remember it as the capital. In conclusion, the statement is false. The capital of Brazil is Brasília, not Rio de Janeiro. Rio de Janeiro was the former capital but is no longer the current capital. You can confidently use the sources listed above to verify this information. |
Débogage et orientation
Vertex AI Studio permet aux développeurs d'accéder aux étapes du raisonnement du modèle dans l'interface utilisateur. Cela est utile lorsqu'ils examinent votre requête en cas de débogage :
Analyser le processus de réflexion
Examinez attentivement le processus de réflexion de Gemini pour identifier les éventuelles erreurs. Cette analyse peut fournir des indices précieux pour affiner vos requêtes et obtenir la réponse attendue de manière cohérente dans votre application.
Exemple de scénario : Si le processus de réflexion de Gemini révèle qu'il a mal compris une étape d'une instruction complexe, vous pouvez affiner la requête pour clarifier cette étape spécifique.
Step 1: Summarize the following article about quantum physics. Step 2: Translate the summary into Spanish. Step 3: Extract the key entities from the Spanish summary. Article: ... [Article Text] ... The user wants me to process an article about Google's new quantum chip called Willow in three steps: Step 1: Summarize the article. Step 2: Translate the summary into Spanish. Step 3: Extract the key entities from the Spanish summary. Let's break down each step. Step 1: Summarize the article. I need to read the article and identify the main points. The article is about Google's new quantum chip, Willow, and its achievements. The key achievements are: Exponential error reduction as qubits scale up. Outperforming supercomputers in a benchmark computation (RCS). Willow is a step towards commercially relevant quantum computing. It's the first system below threshold for quantum error correction. It achieved a benchmark computation in under 5 minutes that would take a supercomputer 10 septillion years. It was fabricated in a new facility. It has state-of-the-art performance and high quality qubits. Google is looking towards real-world applications with Willow. Quantum computing will benefit AI. I will synthesize these points into a concise summary. Step 2: Translate the summary into Spanish. Once I have the summary in English, I will translate it into Spanish. I need to ensure the translation is accurate and natural-sounding in Spanish. Step 3: Extract the key entities from the Spanish summary. After translating the summary, I will identify and extract the key entities. Key entities are typically nouns or noun phrases that represent important objects, concepts, or actors. In this context, they will likely include: Names of chips (Willow, Sycamore) Companies/Organizations (Google Quantum AI, Google Research) Technologies (quantum chip, quantum computer, qubits, quantum error correction, superconducting quantum system, classical computers, supercomputers, AI) Benchmarks (random circuit sampling, RCS) People (Julian Kelly, Peter Shor, David Deutsch, Sergio Boixo, Hartmut Neven, John Preskill) Concepts (exponential error reduction, below threshold, beyond breakeven, state-of-the-art performance, commercially relevant applications, quantum mechanics, multiverse, T1 times, system benchmarks, algorithmic benchmarks) I will need to carefully review the Spanish summary to identify these entities and list them out. |
Provide corrective feedback
If you observe errors or inconsistencies in Gemini's reasoning, provide corrective feedback in your prompt and the system instructions to anticipate where the model's response may deviate from your desired outcome.
You are an expert in botany. When asked to identify a plant, focus on leaf shape, flower structure, and stem characteristics. Identify this plant based on these characteristics: ... [Plant Characteristics Description] ... Previous incorrect response focused on habitat only. Please disregard habitat and focus on morphology. Previous incorrect response focused on habitat only. Please disregard habitat and focus on morphology. Tabebuia vellosoi |
Tirer le meilleur parti des sorties longues et de la réflexion longue
La réflexion génère des "pensées" dans la sortie de jeton, qui présentent le raisonnement détaillé de Gemini. Pour exploiter pleinement ces sorties :
Examiner le raisonnement
Examinez attentivement le processus de raisonnement de Gemini (disponible dans l'interface utilisateur de Vertex AI Studio), la façon dont il a divisé la tâche et les choix qu'il a faits pour votre sortie. Cela peut vous aider à comprendre comment il en est arrivé à cette conclusion.
Fournir des conseils dans le raisonnement
Si vous souhaitez obtenir une sortie particulièrement longue, vous pouvez fournir plus d'instructions dans votre requête afin de limiter la réflexion nécessaire pour accomplir la tâche présentée dans la requête. Vous réservez ainsi une plus grande partie de la sortie de jeton pour votre réponse.
*Exemple : Pour un rapport détaillé, vous pouvez spécifier des sections et des sous-sections dans votre requête afin de guider la réflexion de Gemini et de structurer efficacement la sortie.
Generate a detailed report on the impact of climate change on coastal ecosystems. Report Structure: Introduction Rising Sea Levels 2.1. Impact on Mangroves 2.2. Impact on Coral Reefs Ocean Acidification 3.1. Impact on Shellfish 3.2. Impact on Plankton Extreme Weather Events 4.1. Increased Storm Intensity 4.2. Coastal Erosion Conclusion Provide detailed analysis within each section. <...> Climate Change Impact on Coastal Ecosystems: A Detailed Report 1. Introduction Coastal ecosystems, the dynamic interface between land and sea, are among the most productive and biodiverse environments on Earth. They provide crucial ecosystem services, including coastal protection, fisheries, carbon sequestration, and support for a vast array of species. However, these vital ecosystems are increasingly threatened by the multifaceted impacts of climate change. Driven by anthropogenic greenhouse gas emissions, climate change is manifesting in several interconnected ways that are profoundly altering the physical and chemical conditions of coastal environments. This report will delve into the key impacts of climate change on coastal ecosystems, focusing on rising sea levels, ocean acidification, and extreme weather events, and their specific consequences for critical habitats like mangroves and coral reefs, and key organisms like shellfish and plankton. Understanding these impacts is crucial for developing effective conservation and management strategies to mitigate the damage and enhance the resilience of these invaluable ecosystems. 2. Rising Sea Levels Rising sea levels are a direct and significant consequence of global warming, driven by thermal expansion of seawater and the melting of land ice (glaciers and ice sheets). This phenomenon poses a direct threat to low-lying coastal ecosystems by inundating habitats, altering salinity regimes, and increasing erosion rates. The rate of sea level rise is accelerating, and projections for the 21st century indicate a continued upward trend, with potentially devastating consequences for coastal environments. 2.1. Impact on Mangroves Mangrove forests are unique coastal ecosystems found in tropical and subtropical intertidal zones. They provide numerous benefits, including coastal protection against storms, nursery grounds for fish and invertebrates, and significant carbon sequestration... <...> |
Étape suivante
Essayez d'utiliser un modèle à raisonnement avec notre notebook Colab, ou ouvrez la console Vertex AI et essayez de demander vous-même au modèle de vous répondre.