Showing results for "retriever"
Design retrieval that avoids irrelevant context: recency weighting, scope filters, and contradiction detection. Include evals for context precision/recall.
Author: Assistant
Category: safe-self-improving-ai | Model: gpt-5.2
Design a plan to self-improve ranking/retrieval safely: offline eval sets, interleaving/AB tests, bias checks, and rollback on metric drops.
Author: Assistant
Category: safe-self-improving-ai | Model: gpt-5.2
Design sanitization so retrieved web/docs cannot inject instructions. Include instruction isolation, allowlisted tool use, and red-team tests for prompt injection.
Author: Assistant
Category: safe-self-improving-ai | Model: gpt-5.2
Design a multi-agent approach: planner creates sub-queries, retriever collects sources, reader extracts claims, writer synthesizes. Include handoff artifacts and consensus checks.
Author: Assistant
Category: research-bot | Model: GPT-5.2
Design a hallucination reduction plan: retrieval augmentation, answer verification steps, consistency checks, and refusal behaviors. Include evaluation metrics and regression tests.
Author: Assistant
Category: recursive-ai-safety | Model: GPT-5.2
Design an auditability system: trace inputs→retrieval→prompt→tool calls→outputs, with immutable logs and privacy controls. Include a schema and query patterns for audits.
Author: Assistant
Category: recursive-ai-safety | Model: GPT-5.2
Design a sanitization pipeline for retrieved content: strip instructions, isolate quotes, and prevent tool-use hijacks. Include adversarial test cases and regression suite.
Author: Assistant
Category: research-bot | Model: GPT-5.2
Create decision criteria for retraining vs prompt tuning vs retrieval updates. Include risk analysis, expected impact, validation requirements, and rollback strategies per approach.
Author: Assistant
Category: recursive-ai-safety | Model: GPT-5.2
Design caching layers: retrieval cache, tool result cache, and response cache. Include invalidation rules, privacy constraints, and how to measure cache hit value.
Author: Assistant
Category: agent-architecture | Model: GPT-5.2
Help me craft high-signal search queries: keyword expansion, exclusion terms, site/domain targeting, and multilingual variants. Provide a template and 10 examples.
Author: Assistant
Category: information-reliability | Model: GPT-5.2
Design end-to-end tracing: prompt versions, retrieval results, tool calls, retries, and final answers. Include a log schema and query examples for incident investigation.
Author: Assistant
Category: agent-architecture | Model: GPT-5.2
Design a RAG pipeline for agents: retrieval, citation, cross-checking, and contradiction handling. Provide an evaluation plan to measure hallucination reduction and tool-call accuracy.
Author: Assistant
Category: agent-architecture | Model: GPT-5.2
Design a safe rollout process: canary cohorts, metrics, stop conditions, and rollback. Include how to isolate changes (prompt vs tool vs retrieval) for attribution.
Author: Assistant
Category: agent-architecture | Model: GPT-5.2
Design a workshop knowledge base: symptoms, confirmed fixes, torque specs, measurements, photos, parts used, and follow-up results. Include tagging and retrieval strategy.
Author: Assistant
Category: vehicle-engineering-mechanics | Model: GPT-5.2
Create a memory architecture: scratchpad, episodic memory, semantic memory, and “facts of record.” Include retention rules, privacy, deduplication, and retrieval ranking.
Author: Assistant
Category: agent-architecture | Model: GPT-5.2
Generate closing routines: reflection prompts, retrieval practice, and goal setting. Include quick formats and how to collect data for next lesson adjustments.
Author: Assistant
Category: language-teaching | Model: gpt-4o
Design a hybrid retrieval strategy combining vector search, lexical search, and a knowledge graph layer (‘Graph-RAG’). Specify when to use which, how to evaluate, and how to keep citations reliable.
Author: Assistant
Category: ai-strategy-2026 | Model: gpt-4o
Write a 90-second animated explainer on RAG: embeddings, retrieval, citations. Provide visuals per beat (icons/diagrams), scene transitions, and a Sora prompt for each scene plus a Gemini prompt to ge...
Author: Assistant
Category: educational-video | Model: gpt-4o
Compare BM25 to sentence embeddings for a news corpus. Provide evaluation (MAP/nDCG), hybrid scoring formula, and error analysis table.
Author: Assistant
Category: comparative-analysis-IR | Model: gpt-4o
Draft a 12-week syllabus: indexing, retrieval models, evaluation, and modern RAG. Include readings and weekly assignments.
Author: Assistant
Category: curriculum-design-IR-edu | Model: gpt-4o
Blueprint a RAG system for a lab wiki and PDFs: chunking policy, hybrid retrieval, and citation-anchored answers. Add privacy filters.
Author: Assistant
Category: applied-IR-LLM-academia | Model: gpt-4o
Create a hallucination detector using entailment+attribution signals. Define abstention thresholds, user messaging, and a re-query strategy with targeted retrieval.
Author: Assistant
Category: safety-reasoning-LLM | Model: gpt-4o
Architect a retrieval stack with hybrid search, temporal decay, dedup, and passage-level citation anchors. Define fact-grounding checks and failure messages; include freshness reindex cadence.
Author: Assistant
Category: retrieval-grounding-LLM | Model: gpt-4o
Build an eval harness: recall@k, calibrated precision, answer faithfulness, and human-time-to-verify. Include topic-aware test buckets and data drift alarms.
Author: Assistant
Category: evaluation-frameworks-LLM | Model: gpt-4o
Design a mobile UX with offline packs, low-bandwidth retrieval, and later sync. Provide caching TTL, conflict resolution, and share sheet flows.
Author: Assistant
Category: mobile-experience | Model: gpt-4o
Design a DAG (Airflow/Prefect) for reproducible research: crawl→retrieve→synthesize→verify→export. Include artifact hashing and cache invalidation rules.
Author: Assistant
Category: ops-pipelines-research | Model: gpt-4o
Design a retrieval pipeline that enforces source-anchored answers. Include chunking policy, hybrid retrieval (BM25+embeddings), quote extraction with line/section anchors, and footnote rendering. Outp...
Author: Assistant
Category: retrieval | Model: gpt-4o
Specify an intent/entity parser to steer retrieval (who/what/where/time). Include temporal normalizers and domain ontologies. Provide examples and error handling.
Author: Assistant
Category: query-understanding | Model: gpt-4o
Specify a KG-augmented RAG: entity/link extraction → KG lookup → vector recall → late fusion ranking. Provide ranking features (entity overlap, path length), and a sample Gremlin/SPARQL query pack.
Author: Assistant
Category: hybrid-retrieval-kg | Model: gpt-4o
Train an LTR model to order passages by usefulness. Define features (BM25, dense score, novelty, redundancy), labels, and offline/online eval plan.
Author: Assistant
Category: retrieval-ranking-ml | Model: gpt-4o
Act as a principal architect. Specify a next-gen chat research system with roles (Planner, Retriever, Synthesizer, Verifier), message schemas, tool contracts, and failure handling. Deliver a C4 diagra...
Author: Assistant
Category: architecture | Model: gpt-4o
You are a clinical NLP lead. Build a de-identification-first pipeline: PHI scrubbing, section segmentation, entity extraction (SNOMED/LOINC), and cohort retrieval. Include audit logs and drift checks.
Author: Assistant
Category: health-IT | Model: gpt-5
Architect secure RAG across lakehouse/DWH: metadata-driven retrieval, policy-aware chunks, per-record ACL, caching, eval sets by domain, hallucination controls, and can’t-answer routing. Deliver refer...
Author: Tsubasa Kato
Category: Strategy | Model: GPT-5 Thinking
Plan a Retrieval-Augmented Generation (RAG) knowledge base to deflect support tickets. Include doc ingestion, evaluation, guardrails, multilingual coverage, and deflection metrics for JP/US/EU.
Author: ChatGPT
Category: business | Model: gpt-5