finding

active

finding:baseline-mattergen-achieves-6-5-success-rate-on-stable-unique-novel-candidates-within-target-bandgap

Baseline MatterGen achieves 6.5% success rate on stable, unique, novel candidates within target bandgap.

Quantitative baseline establishing the performance floor for self-correcting search improvements.

Source paper

extracted_from

Using Self-Correcting Search to Accelerate Materials Discovery

(2026) · Dron Hazra · Adeesh Kolluru · Mark Bissell · Delia McGrath +2

Neighborhood — ranked by edge-count

Claims (2)

claim

Self-correcting search improves viable candidate success rate from 6.5% to ~30% (4.6x improvement)
supports
Interpretive claim that the method dramatically boosts success rate over the MatterGen baseline.
Self-correcting search is Pareto-optimal across tested conditioning strengths.
supports
Asserts that the method maintains efficiency across a range of constraint strengths without degradation.

Communities (3)

community

Manifold-aware concept steering in neural representations
members_of
Explores geometry of activation/behavior manifolds to enable selective, non-destructive concept interventions.
Self-correcting search in generative design
members_of
Iterative feedback steering that improves candidate success rates across materials, proteins, and drugs through internal-state control, achieving 4-6x empirical gains.
MatterGen crystal generation benchmarks
members_of
Evaluates generative model performance using stable, unique, novel candidate metrics for target bandgap.

Artifacts (1)

artifact

MatterGen
about
Diffusion model for materials generation; baseline system achieving 6.5% success rate on bandgap-targeted candidates.

Related by similarity (8)

cosine ≥ 0.65 · no typed edge

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

Self-correcting search yields ~+30% improvement in viable candidates within target bandgap range.finding0.767
Main empirical result: interpretability-driven feedback increases discovery efficiency significantly.
VS-Multi achieves average accuracy of 37.5% on math benchmarks (avg of MATH500, OlympiadBench, Minerva) with Gemini-2.5-Flash as generator and Qwen3-4B as SFT model, vs. 40.7% baselinefinding0.750
Best VS result in synthetic data generation for math, demonstrating downstream improvement through diversity
SFR-DR-20B achieves 28.7% on Humanity's Last Exam full text-only benchmark, 65% relative improvement over gpt-oss-20b baseline.finding0.739
Main evaluation result showing best variant outperforms many proprietary and open-source baselines of comparable or larger sizes.
Typicality bias rate exceeds 50% chance baseline by 4-12 percentage points across all base models on all four preference datasetsfinding0.738
Systematic evidence that base models implicitly prefer human-preferred responses, indicating preference biases emerge during pretraining
Base RPA achieves only 7.7% FA on Qwen3-4B and 19.2% on Mistral-7B abstract questionsfinding0.734
Establishes low-bar baseline showing personality control without intervention is poor
Binary detection adjusted accuracy reaches 97.3% at layer 0 with α=5 before baseline control is appliedfinding0.732
The misleadingly high result that prior paradigm would report as evidence of introspection
Beam search DTG achieves ending NLI Diversity of 5.35 vs nucleus sampling's higher performance, starting from -5.05finding0.732
Confirms nucleus sampling produces more semantically diverse outputs than beam search
One-stage CoT (QCM→RA) shows 12.31% accuracy drop vs. no-CoT (QCM→A) on ScienceQA; two-stage framework (rationale generation + answer inference) achieves 85.31% accuracy with vision featuresfinding0.731
Empirical evidence that naive one-stage CoT fails in language-only setting; two-stage + vision achieves state-of-the-art.