Showing results for "RAG"
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
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 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 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
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
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
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
Design a workflow where content drafts are generated from a curated knowledge base with citations. Output: ingestion rules, citation format, hallucination checks, and an authoring template that forces...
Author: Assistant
Category: content-generation-2026 | 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
Design a provenance layer for internal RAG: citation IDs, document signing, version pinning, and ‘evidence packs’ in responses. Include UX and evaluation methods.
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
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
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
Design an evaluation harness that measures relevance and faithfulness for IR+LLM answers. Include human labeling rubric and inter-rater checks.
Author: Assistant
Category: eval-framework-IR-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
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
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
Act as an NLP architect. Build a KG from UMLS/MeSH and a RAG stack for question answering: entity linking, relation extraction, indexing strategy, and evaluation (exact match, factuality).
Author: Assistant
Category: biomed-NLP | Model: gpt-5
You are a metrics lead. Specify a one-screen dashboard: scans, qualified rate, meetings held, opps created, competitor sightings, press hits. Include RAG thresholds and stand-up prompts.
Author: Assistant
Category: trade-show | Model: gpt-4o
Design a data-lite RAG plan without engineers. Sources: Google Drive/Notion/PDFs. Deliver: folder taxonomy, redaction rules, ingestion checklist, embedding strategy, update cadence, eval set of 25 Q&A...
Author: Tsubasa Kato
Category: Strategy | Model: GPT-5 Thinking
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
Act as an AI rollout lead for a sub-50 person company. Deliver a 14-day plan: use cases ranked by ROI/risk, 2 quick-win agents (inbox triage, FAQ/RAG), minimal governance (human-in-the-loop), success ...
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