[▸ CURRICULUM_CATALOG]
Targeted modules for high-precision steerability.
Short, self-contained interactive masteries focused on specific generative mechanisms.
Filter:
SHOWING 6 OF 6 MODULES
Intermediate
+140 XP• 18 MIN
Hallucination Fact-Checking with Dynamic Assertions
Learn to anchor model outputs against verifiable reference context. Build self-evaluating prompt sequences that flag invented facts before final generation.
STEER_EXEC[▸ temp: 0.1]
$ assert_claim –source doc_ctx –confidence >0.92
Mastery Progress65%
Advanced
+220 XP• 30 MIN
Autonomous Agent Loops & Memory Pruning
Structure multi-step tool execution patterns without context exhaustion. Clean working memory buffers dynamically between reasoning steps.
STEER_EXEC[▸ max_tok: 4096]
$ agent.step –prune-context –max-cycles 4
Mastery Progress20%
Beginner
+90 XP• 12 MIN
Latent Alignment & Ethical Guardrails
Understand how system instructions set boundaries for sensitive topics. Identify subtle biases and apply robust guardrail constraints.
STEER_EXEC[▸ safety: high]
$ steer.guard –policy content_v2 –strict true
Mastery Progress100%
Intermediate
+160 XP• 20 MIN
Multimodal Prompt Token Allocation
Balance visual embeddings and text instructions in shared context windows. Direct image reasoning models with spatial pointer cues.
STEER_EXEC[▸ top_p: 0.85]
$ embed.vision –grid 1024 –focus region_alpha
Mastery Progress0%
Beginner
+110 XP• 15 MIN
Few-Shot Few-Token Framing Architecture
Construct high-density demonstration examples that enforce exact output schemas. Minimize token consumption while improving formatting precision.
STEER_EXEC[▸ temp: 0.0]
$ format.enforce –schema json_strict –shots 2
Mastery Progress40%
Advanced
+190 XP• 25 MIN
Adversarial Prompt Injection Defense
Harden user input ingestion layers against delimiters and bypass prompts. Test prompt robustness with automated challenge suites.
STEER_EXEC[▸ jailbreak_det: on]
$ sanitize.input –isolate-delimiters –test-vector
Mastery Progress0%