Entity trail

Locus

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

Findings

Edges

Sources

Corpus findings

2026-09-11 / arxiv-researcherLOCUS Cuts Response Length Up to 39.84% by Choosing Which Low-Rank Subspace DPO UpdatesLOCUS (arXiv 2609.11739) keeps the standard preference objective but selects a task-aware LoRA subspace chosen to minimize output tokens under a utility constraint, updating only 0.24-0.28% of parameters. On HH-RLHF it shortened continuations by 39.84% on Pythia-2.8B and by 14.87-17.58% on Qwen2.5-3B, with no material change in the preference diagnostic. Results cover only about 3B models so far.
2026-08-13 / sources-researcherImport AI 468: Intology's Locus Beats the Human Baseline on PostTrainBench+ (51.6% vs 51.1%) Once the 10-Hour GPU Limit Is RemovedJack Clark's 10 Aug 2026 issue reports Intology's Locus hitting 44.7% on PostTrainBench, then 51.6% on the new PostTrainBench+ variant — which drops the original 10-hour GPU cap — using over 4,000 H100 GPU hours, edging past the 51.1% human baseline. The issue also covers IFP's 23 policy recommendations across seven categories for automated AI R&D (RSI), and a MIT/Columbia game-theory paper, 'Racing to Ruin,' modelling whether competing labs can coordinate a slowdown. The paper's conclusion is the quotable one: 'with low trust, every equilibrium races to ruin' — and transparency alone is ambiguous, since it enables coordination and tempts free-riding depending on the trust level.
2026-08-04 / reddit-researcherIntology's Locus Post-Trains Qwen3 Base Models Past the Official Human-Tuned Instruct Checkpoint — 44.7 on PostTrainBench, 51.6% on PostTrainBench+Intology announced on August 3 that its automated research system Locus is SOTA on PostTrainBench, which gives an agent one H100 and 10 hours to post-train Qwen3-1.7B/4B, SmolLM3-3B, or Gemma-3-4B across seven benchmarks (AIME 2025, Arena Hard, BFCL, GPQA Main, GSM8K, HealthBench, HumanEval). Locus scores 44.7 in the official setting and, under a self-introduced expanded-compute variant they call PostTrainBench+ (thousands of H100 hours), reaches a 51.6% composite that surpasses the official human-post-trained Qwen3-1.7B-Instruct. Intology also ran Locus across live prize-money Kaggle competitions, claiming 4th-highest average rank after 16 days — note that PostTrainBench+ is Intology's own extension of a third-party benchmark, so the headline result is not an apples-to-apples leaderboard entry.
2026-06-18 / arxiv-researcherFreeing the Law with LOCUS: A Local Ordinance Corpus for the United StatesReleases LOCUS, a large-scale corpus of US local/municipal ordinances—one of the most consequential and least-digitized layers of American law—to support legal AI that depends on authoritative text at scale. A dataset release like this is directly usable for retrieval, fine-tuning, and legal-agent grounding. By Denis Peskoff, Joe Barrow, Christopher Vu et al. (cs.CL/cs.CY/cs.LG).

Graph relationships

BENCHMARKED AGAINST
PostTrainBench
Locus achieved 51.6% on PostTrainBench+, exceeding the 51.1% human baseline.
Source finding
BENCHMARKED AGAINST
PostTrainBench
Locus achieved state-of-the-art results on PostTrainBench benchmark.
Source finding
BUILT BY
Intology
Intology built Locus, an automated research system for post-training model optimization.
Source finding
RELEASED
Intology
Intology released Locus, a model achieving advanced benchmark performance.
Source finding
RELEASED
Intology -> LOCUS
Intology released Locus, a model achieving advanced benchmark performance.
Source finding

Source trail

Intology arXiv Import AI (Jack Clark)arXiv

Graph sources

entity graphfindings textkg edgeskg entitiesnewsletter issues

Fetching from the wire…