Search

What are you looking for?

Search our services, use cases and practical insights.

Enter at least 2 characters

Popular starting points

AI explained simply

Alignment and human feedback in AI

RLHF stands for reinforcement learning from human feedback. It uses human evaluations as a training signal to further optimise a language model's behaviour. A pretrained language model can generate language, but this alone does not reliably teach it which answers people consider helpful, appropriate or safe. Alignment covers methods and controls that bring its behaviour closer to such expectations. RLHF is one training method within this broader effort – not the complete solution.

The short answer

Alignment is the overarching goal of bringing an AI system's behaviour closer to its intended tasks, human instructions and safety rules. In reinforcement learning from human feedback, or RLHF, people compare model responses. Their preferences are converted into a reward signal that is used to optimise the language model further.

In brief

  • Alignment is an overarching goal, not the name of a single training method.
  • Classical RLHF uses human comparisons, a learned reward model and reinforcement learning.
  • Supervised fine-tuning, RLHF and human approval during live operation serve different purposes.
  • RLHF guarantees neither truth nor complete safety and does not replace company-specific controls.

Alignment is a goal, not a single method

During pretraining, a language model is primarily optimised to generate likely continuations. Reliably following instructions, refusing inappropriate requests or observing the rules of a particular company does not emerge automatically. Alignment attempts to bring observable behaviour closer to such expectations.

There is no single universal measure of a fully aligned model. People may judge instructions, risks and values differently. It therefore matters whose requirements a system was aligned to and for which use its behaviour has actually been tested.

Model alignment
Training and evaluation shape general behaviour such as instruction following, handling uncertainty and appropriate refusals.
Application alignment
System instructions, approved sources, tools, permissions and guardrails adapt a general model to a specific use case.
Operational alignment
Ownership, evals, monitoring, escalation and human decisions keep the application within its intended limits during everyday use.

Pretraining, fine-tuning, RLHF and human review serve different purposes

Some of these methods build on one another, but they should not be treated as equivalent. A classical RLHF process often starts with a model that has already been pretrained and further optimised using demonstrations. A person reviewing a live business response is not performing model training.

Pretraining, fine-tuning, RLHF and human review serve different purposes
MethodLearning or review signalPrimary purposeDoes not replace
PretrainingVery large text and data collections, usually self-supervised learningDevelop general language and pattern capabilitiesReliable instruction following and deployment rules
Supervised fine-tuningCurated inputs paired with target responsesLearn tasks, formats or desired responses from demonstrationsCurrent knowledge, permissions and ongoing quality control
Reinforcement learning from human feedback (RLHF)Human comparisons or ratings of model responsesOptimise behaviour towards the evaluated preferencesFactual accuracy, universal values and use-case accountability
Human review in operationA responsible person assesses a specific real caseApprove, correct or escalate before an outcome takes effectDoes not automatically change the model and is not RLHF

How classical RLHF works

RLHF stands for reinforcement learning from human feedback. Rather than expressing every quality of a desired response as a mathematical rule, people indicate which of several responses they prefer. A classical process first uses those comparisons to train a reward model and then uses that model to optimise the language model.

  1. Step 1

    Define the desired behaviour

    Guidelines and examples describe what should count as helpful, appropriately uncertain or unacceptable.

  2. Step 2

    Generate response alternatives

    The model produces several possible responses to the same prompts.

  3. Step 3

    Ask people to compare them

    Trained raters score or rank alternatives according to a defined rubric.

  4. Step 4

    Model the preferences

    In classical RLHF, a reward model learns to predict which responses raters would prefer.

  5. Step 5

    Optimise the language model

    Reinforcement learning changes the model policy so responses with a higher predicted reward become more likely.

  6. Step 6

    Evaluate independently

    Separate cases examine improvements, side effects, excessive refusals and new failure modes.

How RLHF relates to reinforcement learning

Reinforcement learning is a broad area of machine learning. A system tries actions and learns from rewards which strategy performs better over time. The reward may come from a simulation, a game score or a technically defined objective.

In RLHF, the objective signal is instead derived indirectly from human preferences. For language models, people typically compare complete responses rather than evaluating each token. The learned signal is then used for reinforcement learning.

RL is the broader field
Reinforcement learning does not require human ratings. An environment can provide rewards automatically.
RLHF is a particular application
Human comparisons help translate qualities that are difficult to specify, such as helpfulness or appropriate behaviour, into an optimisation signal.
Not all preference training is RLHF
Methods such as direct preference optimisation learn directly from preference data without a separate reward model and reinforcement learning. Feedback supplied by AI systems rather than people is often called RLAIF.

Where model alignment ends and company responsibility begins

A provider can train a general model for helpfulness and broad safety requirements. This does not teach it the binding processes or risk thresholds of an individual company. Productive use therefore requires evaluation of the entire system, not only the description of its base model.

  • Run independent evals with representative and critical company cases
  • Technically limit approved sources, tools, actions and permissions
  • Require responsible people to approve binding decisions and exceptions
  • Retest model and provider updates before rollout
  • Do not turn feedback into training data without review and data governance
Preferences are not objective
Ratings depend on instructions, expertise, language, culture and the composition of the rater group. Minority views or rare cases can be underrepresented.
Rewards are proxies
A reward model represents human expectations imperfectly. Excessive optimisation can satisfy the measurement signal without reliably achieving the underlying goal.
Desired behaviour is not the same as truth
A persuasively phrased response can still be factually wrong. RLHF is neither a knowledge base nor proof that information is current.
New situations remain difficult
Behaviour that works on training examples can fail for other languages, specialist domains, attacks or unusual inputs.
Example from day-to-day business

Example: a customer service assistant with explicit limits

A company uses a language model that its provider has already aligned for helpful responses using demonstrations and preference data. This is not sufficient for the company's own service process: the assistant may use only approved product information, must not make binding contractual commitments and cannot issue goodwill payments on its own. Such cases are handed to employees. A fixed eval set checks source use, appropriate refusals and escalation after each model update. General model alignment, application-specific rules and human review complement one another rather than replace one another.

What to remember

Alignment is a layered process. RLHF can improve general model behaviour, but company-specific rules, independent evals, technical limits and human accountability are still required for controlled business use.

Would you like to apply this to your situation?

Together, we clarify what makes sense for your process, data and systems – in plain language and without unnecessary complexity.

Discuss Your Project