Results for “pillar-alignment”

Scores generated image captions against reference captions and source images using the Polos metric, which is trained to align with human judgments and probes hallucination robustness and open-vocabulary evaluation.

Clade goal-driven autonomous improvement loop (Blueprint architecture — deterministic pre/post phases + LLM supervisor/worker nodes, converges when goal met or max-iter hit). NOT the Claude Code built-in /loop (which polls a prompt on an interval like `/loop 5m /foo`) — if the user wants interval polling, route to the built-in.

Rewrites free-form draft prompts into the conventions Claude Opus 4.6 or 4.8 responds best to, either inline or by overwriting a draft file.

Perform nonlinear dimensionality reduction, 2D/3D embeddings, clustering preprocessing, supervised or semi-supervised UMAP, DensMAP, AlignedUMAP, and Parametric UMAP workflows using the umap-learn library.

Evaluates automated metrics and vision-language models on identifying granular errors in detailed image descriptions and ranking paired descriptions against human judgments, using macro F1, pairwise accuracy, Spearman rank ρ, and Kendall's τ.

Model a Phase-Locked Loop (PLL) IC from its datasheet or system specs using Mixed-Signal Blockset. Without this skill, agents universally select the wrong solver and produce non-functional PLL models — 100% of unguided attempts fail. Covers Integer-N, Fractional-N, Dual Modulus architectures, loop filter design, lock time optimization, VCO phase noise configuration, and msbPllArchitectures/msbPllFoundation block assembly. Use when: PLL modeling, frequency synthesizer design, phase noise simulation, lock time analysis, charge pump design, loop filter tuning, datasheet-to-model, Mixed-Signal Blockset PLL, msbPllArchitectures.

Trains LLMs with SimPO, a reference-free preference optimization method that outperforms DPO, using configurable hyperparameters and workflows for various models and tasks.

Pixtral 12B: A Frontier Multimodal Model

Simple Preference Optimization for LLM alignment. Reference-free alternative to DPO with better performance (+6.4 points on AlpacaEval 2.0). No reference model needed, more efficient than DPO. Use for preference alignment when want simpler, faster training than DPO/PPO.

Co-fold biomolecular complexes with Boltz-2, an open AlphaFold3-style model — predict protein + ligand (SMILES/CCD), protein + nucleic-acid, and multi-chain structures in one pass, with binding-affinity prediction. Use when folding a protein together with a small-molecule ligand, predicting a holo (ligand-bound) complex or its binding affinity, or co-folding protein–DNA/RNA assemblies. For protein-only or protein–protein folding without ligands prefer alterlab-alphafold; for antibody–antigen complexes prefer alterlab-chai; to dock a ligand into a FIXED receptor structure prefer alterlab-diffdock; to look up an existing structure prefer alterlab-pdb. Part of the AlterLab Academic Skills suite.

The two model-touching halves of Worklink, shipped as one opt-in skill. (1) Planner: decompose a parent Chainlink issue into testable worklink:ready leaf subissues (mutates Chainlink only, never executes). (2) Ready-queue poller: discovers worklink:ready leaves and dispatches them by invoking `mimir worklink run` as a detached subprocess, up to the concurrent-claim cap — it never reimplements claim/evidence/transition. Opt-in (mimirbot yes, muninn no): `mimir skills install chainlink-orchestrator`, then set the env below to enable autonomous dispatch.

Simple Preference Optimization for LLM alignment. Reference-free alternative to DPO with better performance (+6.4 points on AlpacaEval 2.0). No reference model needed, more efficient than DPO. Use for preference alignment when want simpler, faster training than DPO/PPO.

Simple Preference Optimization for LLM alignment. Reference-free alternative to DPO with better performance (+6.4 points on AlpacaEval 2.0). No reference model needed, more efficient than DPO. Use for preference alignment when want simpler, faster training than DPO/PPO.

Simple Preference Optimization for LLM alignment. Reference-free alternative to DPO with better performance (+6.4 points on AlpacaEval 2.0). No reference model needed, more efficient than DPO. Use for preference alignment when want simpler, faster training than DPO/PPO.