Practical insights
- 1000published
- 25in review
- 8topic families
Results
- Estimate key-value cache memory before promising long contexts on self-hosted models
An ML engineer must serve a self-hosted model with long contexts to many concurrent users on a fixed GPU budget.
harddocumented fact - Consider continued pretraining only when you hold a large unlabelled domain corpus
An insurer's claims text is full of domain jargon, and a general model handles it poorly even after prompt work.
hardempirical finding - Expect LoRA to learn less and forget less than full fine-tuning
A team uses LoRA to teach a model a large new skill domain, such as a proprietary programming language, and gains are small.
hardempirical finding - Locate RLHF failures in feedback, reward model or policy before changing the recipe
After preference tuning, an internal assistant becomes evasive on some topics and overconfident on others, and the team considers switching alignment methods.
hardpractitioner heuristic - Track a gold metric while optimising against a reward model, and stop before it turns down
During RL fine-tuning, the reward-model score keeps rising run after run, and the team keeps training.
hardempirical finding - Choose DPO for simplicity, but check for distribution shift in the preference data
A team with a fixed preference dataset picks DPO because it avoids training a reward model and running reinforcement learning.
hardempirical finding - Decompose multi-hop questions instead of hoping one retrieval finds everything
Questions combine facts from different documents, such as "which of our A321 routes are affected by the new crew-rest rule?"
hardpractitioner heuristic - Treat writes to long-term agent memory as untrusted input with provenance
An assistant saves "facts" or preferences from conversations, documents or web pages into persistent memory reused in later sessions.
hardpractitioner heuristic - Close exfiltration channels such as auto-loaded links and images in model output
The chat interface renders model output as Markdown, including images and links to arbitrary domains, and the model can see private data.
hardpractitioner heuristic - Keep the model that reads untrusted content away from tools and private data
One agent reads incoming customer emails, can query customer records, and can send messages externally.
harddocumented fact - Fix the plan before reading untrusted data when the task allows it
An agent must process external content, for example summarise today's supplier emails into a report, with a predictable set of actions.
harddocumented fact - Have the model quote the relevant evidence before it answers from a long input
Retrieval is perfect in a 40,000-token prompt, yet multi-step answers over the supplied evidence are worse than on a trimmed version of the same prompt.
hardempirical finding - Re-check preconditions at write time with a version token, not only at read time
An agent reads seat availability or a fare early in a long run and makes the booking change many steps later.
harddocumented fact - Record the undo step for every write and place irreversible steps last
An agent must perform a sequence of writes, such as hold a seat, charge a fee and send a confirmation, and a later step may fail.
harddocumented fact - Run a new write-capable agent in shadow mode and compare its proposals with human actions first
A team wants to give a drafting agent permission to apply refunds or rebookings after good offline test results.
hardpractitioner heuristic - Assert required steps and forbidden actions in a trajectory instead of matching an exact path
An agent that issues refunds reaches the right final state, but sometimes skips the identity check the policy requires.
hardpractitioner heuristic - Report code-generation pass@k from n samples with the unbiased estimator, at the k users actually get
A code or SQL assistant is scored by "any of k samples passes the tests", using k samples per task exactly.
harddocumented fact - Gate on a non-inferiority margin for the paired difference, not on a raw score threshold
A CI gate fails the build when the candidate scores below the baseline by any amount on a 400-item set.
hardpractitioner heuristic - Control who can write to the corpus, because a few planted passages can steer answers
A RAG assistant indexes a wiki, a ticket system and a shared drive that many employees and contractors can edit.
hardpractitioner heuristic - Restrict tool access for the rest of a session once untrusted text has entered the context
One long agent session reads external content such as email, web pages or tickets and also holds state-changing tools.
hardpractitioner heuristic - Measure guardrail false positives on benign traffic, because rare attacks make most blocks wrong
A classifier blocks prompts it believes are attacks or policy violations, and its quality is reported as recall on a balanced test set.
hardpractitioner heuristic - Do not extrapolate small-model 8-bit results to large models, because outlier features break naive quantisation
A model that quantised cleanly at a few billion parameters is scheduled for the same naive 8-bit recipe at a much larger size.
hardempirical finding - Reserve a small random slice of routed traffic for the other model so you can measure what routing hides
A router sends easy requests to a small model and the team reports quality from the traffic the router chose.
hardpractitioner heuristic - Isolate long-prompt traffic or chunk prefill when document requests stall chat token streams
A shared deployment serves short chat turns and occasional 50,000-token document analyses, and chat users report streams that freeze for seconds.
hardempirical finding - Size the fallback model for the whole failover load and shed low-priority traffic first
The primary provider degrades and the fallback chain sends 100% of traffic to a secondary model whose quota was set for occasional use.
hardpractitioner heuristic