Skip to content
OpenTrain AIFor AI Companies

KTO: Model Alignment as Prospect Theoretic Optimization

Published Feb 1, 2024
arXiv PDF
Researcher verdict
Starting point
Use as implementation starting point
Benchmark evidence
Thin evidence
Verify before relying
Time to first repro
A few days
Plan setup time
Risk flags
2
Review before use

Results and benchmarks

Freshness tier: hot
KTO: Model Alignment as Prospect Theoretic Optimization is the primary contribution described in this paper.
Task Dataset Metric Value Source
Model Alignment As Prospect Theoretic Optimization MMLU EM 57.2 paper-derived
Model Alignment As Prospect Theoretic Optimization HumanEval EM 54.4 paper-derived

Audit each benchmark finding before selecting an implementation path. Evidence refs map to the disclosure below.

Implementation

Best maintained implementation now

Recommended
Confidence: High
Reproducibility: Limited
contextualai/halos

A library with extensible implementations of DPO, KTO, PPO, ORPO, and other human-aware loss functions (HALOs).

910 stars · 51 forks · Last push Sep 30, 2025 · Apache-2.0 license

  • License
  • CI
  • Dependencies
  • Docker

Official implementation from Papers with Code · Repository link is mentioned in the paper metadata · Community adoption signal (910 stars)

Why this implementation
Confidence: high

contextualai/halos is the strongest maintained implementation based on ranking signals. License is declared (Apache-2.0).

Open contextualai/halos
Reproduction risks
  • No CI workflows detected
  • Dependency manifest is missing
  • Selected contextualai/halos as the strongest maintained implementation for new work.
  • Repository activity is within the last 24 months.

Compare implementation paths

Compare maintenance quality, reproducibility coverage, and evidence confidence before choosing a reproduction baseline.

contextualai/halos
best maintained
Maintenance
Stale risk
Confidence
High
Reproducibility
Limited
Stars
910
Last push
Sep 30, 2025 (330d)

Official implementation from Papers with Code · Repository link is mentioned in the paper metadata

  • No CI pipeline detected
  • No tagged releases
  • No Docker setup
tidedra/vl-rlhf
alternative
Maintenance
Stale
Confidence
Low
Reproducibility
Moderate
Stars
202
Last push
Nov 15, 2024 (649d)

Community adoption signal (202 stars)

  • No push in 12+ months
  • No CI pipeline detected
  • No tagged releases
openbmb/eurus
alternative
Maintenance
Stale
Confidence
Low
Reproducibility
Limited
Stars
323
Last push
Sep 18, 2024 (707d)

Community adoption signal (323 stars)

  • No push in 12+ months
  • No CI pipeline detected
  • No tagged releases

Reproduction readiness

Time to first repro: days
Last checked: Aug 25, 2026

Major work

No dependency manifest, manual reconstruction required

  • contextualai/halos has no requirements.txt, environment.yml, pyproject.toml, or Dockerfile.
  • You will need to reverse-engineer dependencies from import statements in the source code.
  • Last push was 330 days ago.
Open contextualai/halos

Hardware requirements

  • Expect multi-day setup/compute for meaningful reproduction based on current guidance.

Repositories and ecosystem

No additional verified repositories beyond the primary recommendation.

Hugging Face artifacts

No trustworthy direct or curated related Hugging Face artifacts were found yet. Use targeted searches to quickly locate candidate models, datasets, and demos.

Search models Search datasets Search spaces

Tip: start with models, then check datasets and spaces if you need evaluation data or demos.

Research context

Tasks

Model Alignment As Prospect Theoretic Optimization

Methods

None detected

Domains

None detected

Evaluation and human feedback data

Open this paper in HFEPX to review benchmark signals, evaluation modes, and human-feedback protocol context.

Open in HFEPX
Explore similar papers

Jump to Paper2Code search queries derived from this paper's research context.

Model Alignment As Prospect Theoretic Optimization

Data includes links from Papers with Code ( CC-BY-SA-4.0 ).