code wiki / llm

topic: llm

25 modules sharing the llm name family (derived from the tree's prefix discipline).

The 'llm' topic family in the Nishi sovereign ecosystem focuses on ensuring the correctness, performance, and reliability of large language models within a modular, decentralized architecture. Modules like nx_llm_arch_census track architectural coverage and benchmarking, while nx_llm_argmax_probe validates pure-forward correctness without sampling, and nx_llm_capcheck assesses deep capabilities of specific models. These modules collectively support rigorous validation, optimization, and deployment of LLMs across the ecosystem.

auto-narrated by the local model from this topic's module headers; links verified against the wiki index.

narrated overview -- maintained by the narration lane, module links verified against this wiki.

moduledescriptionlinesfuncs
nx_llm_arch_census.nxSOVEREIGN LLM-ARCHITECTURE COVERAGE CENSUS (July 2026 leaderboard). Grounded in756
nx_llm_argmax_probe.nxPURE-FORWARD correctness test. Bypasses run_v3's sampler + repetition-penalty:933
nx_llm_batch_serve.nxthe SOVEREIGN CONTINUOUS-BATCHING LLM SEAT.28810
nx_llm_capcheck.nxSOVEREIGN DEEP CAPABILITY CHECKER for the no-float Qwen2.5-0.5B (our own "Unsloth-style"1505
nx_llm_diff_probe.nxNUMERICAL cross-check vs the transformers oracle (same GGUF). Dumps our forward's1127
nx_llm_forward_profile.nxclean per-decode-forward timing on the1054
nx_llm_france_probe.nxDECISIVE correctness probe (post QKV-bias fix). Feeds the REAL factual prompt753
nx_llm_gap_registry.nxwhere the team needs an LLM to INTERCEDE, marked so a future CUSTOM LLM666
nx_llm_gap_registry_test.nxthe team records WHERE it needs an LLM (this session's real gaps),543
nx_llm_landscape_research.nxTHIN structured research organ for the NEXT-STEPS decision: which small271
nx_llm_layerlens_probe.nxSOVEREIGN logit-lens bisection (no external oracle). Prefills "The capital of1154
nx_llm_loopb_gate.nxTHE LOOP-B CLOSURE: load OUR sovereignly-trained model (nx_f32_qwen2_train_gate's1034
nx_llm_ppl_bench_gate.nxTHE QUALITY BENCHMARK: perplexity + next-token accuracy of ANY gguf model on a1429
nx_llm_probe.nxone-shot probe of the LIVE sovereign LLM seat (127.0.0.1:11434) to MEASURE whether the765
nx_llm_provider.nxa BACKUP to Claude for the LLM-rung work (operator: "figure out a plan to have869
nx_llm_provider_local_test.nxthe identified LOCAL backup is preferred over paid APIs. With a363
nx_llm_provider_test.nxthe team has a BACKUP to Claude for the LLM rung and prefers sovereignty.403
nx_llm_run.nxend-to-end LLM inference scaffold.2773
nx_llm_run_v2.nxend-to-end Llama-class inference composer (real2172
nx_llm_run_v2_test.nxend-to-end smoke for nx_llm_run_v2.nx.2433
nx_llm_sched.nxsovereign CONTINUOUS-BATCHING scheduler (Orca / vLLM1937
nx_llm_sched_gate.nxMEASURED gate for the CONTINUOUS-BATCHING2285
nx_llm_train_estimate.nx666
nx_llm_train_setup_gate.nx773
nx_llm_wired_expand_gate.nx955