Results for "user modeling"
Inferring human goals from behavior.
Running new model alongside production without user impact.
Incrementally deploying new models to reduce risk.
A high-priority instruction layer setting overarching behavior constraints for a chat model.
Letting an LLM call external functions/APIs to fetch data, compute, or take actions, improving reliability.
Controlled experiment comparing variants by random assignment to estimate causal effects of changes.
Requirement to inform users about AI use.
Startup latency for services.
Modeling environment evolution in latent space.
A mismatch between training and deployment data distributions that can degrade model performance.
A continuous vector encoding of an item (word, image, user) such that semantic similarity corresponds to geometric closeness.
The text (and possibly other modalities) given to an LLM to condition its output behavior.
Fine-tuning on (prompt, response) pairs to align a model with instruction-following behaviors.
A preference-based training method optimizing policies directly from pairwise comparisons without explicit RL loops.
Automated detection/prevention of disallowed outputs (toxicity, self-harm, illegal instruction, etc.).
Selecting the most informative samples to label (e.g., uncertainty sampling) to reduce labeling cost.
Time from request to response; critical for real-time inference and UX.
Attacks that infer whether specific records were in training data, or reconstruct sensitive training examples.
Tendency to trust automated suggestions even when incorrect; mitigated by UI design, training, and checks.
Logged record of model inputs, outputs, and decisions.
Methods for breaking goals into steps; can be classical (A*, STRIPS) or LLM-driven with tool calls.
Legal or policy requirement to explain AI decisions.
Recovering training data from gradients.
Inferring sensitive features of training data.
Maintaining two environments for instant rollback.
Shift in feature distribution over time.
Agent calls external tools dynamically.
Cost to run models in production.
Models accessible only via service APIs.
Assigning a role or identity to the model.