simpo-training

simpo-training is a skill for Claude Code, Codex from MilkyWay008/Hermes-OTG. It costs 14 tokens per session (1,631 once invoked), scanned A, a copy of simpo, MIT.

A method for teaching a language model to prefer better answers using chosen and rejected examples, without a second reference model.

In plain words
What is it for?
Use it to align models such as Mistral 7B with preference datasets through SimPO training scripts.
Why use it?
It simplifies preference training by removing the need to run and maintain a separate reference model.

Skill for Claude CodeCodex

Written for no agent in particular: nothing here depends on one.

Needs its repository: it runs a file that does not travel with it, so clone the repository first. The line is output_dir: ./outputs/mistral-7b-simpo.

Good fit Use it to align models such as Mistral 7B with preference datasets through SimPO training scripts.

Compare 6 skills from other repositories ↓
Install

Getting it into your agent

It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.

Clone the repo
git clone --depth 1 https://github.com/MilkyWay008/Hermes-OTG
agentmods
npx agentmods add skills/milkyway008/hermes-otg/simpo

Made for: Claude Code, Codex.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for simpo-training

README.md
[![agentmods](https://agentmods.dev/badge/skills/milkyway008/hermes-otg/simpo/github.svg)](https://agentmods.dev/skills/milkyway008/hermes-otg/simpo)
Your own site
<a href="https://agentmods.dev/skills/milkyway008/hermes-otg/simpo"><img src="https://agentmods.dev/badge/skills/milkyway008/hermes-otg/simpo/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for simpo-training

Your own site · 80×15
<a href="https://agentmods.dev/skills/milkyway008/hermes-otg/simpo"><img src="https://agentmods.dev/badge/skills/milkyway008/hermes-otg/simpo.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 14 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,631 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin 98% copy Near-identical to another mod in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00014 $0.01631
Opus 5 $0.00007 $0.00816
Sonnet 5 $0.00003 $0.00326
Haiku 4.5 $0.00001 $0.00163

Measured 7d ago against content hash 4ebaa42679f7, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-10, from the pricing page.

Security

Grade A, and why

simpo-training scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 7d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

Origin

This is a copy

98% identical to simpo — 2 lines differ, which has more behind it and is treated as the original. This page carries a canonical link to it rather than competing with it.

data/skills/mlops/simpo/SKILL.md · 224 lines

How it starts

The opening of the file, as written. The whole thing — 224 lines — stays where its author put it; the contents beside it link to each section on GitHub.

SimPO - Simple Preference Optimization

Quick start

SimPO is a reference-free preference optimization method that outperforms DPO without needing a reference model.

Installation:

# Create environment
conda create -n simpo python=3.10 && conda activate simpo

# Install PyTorch 2.2.2
# Visit: https://pytorch.org/get-started/locally/

# Install alignment-handbook
git clone https://github.com/huggingface/alignment-handbook.git
cd alignment-handbook
python -m pip install .

# Install Flash Attention 2
python -m pip install flash-attn --no-build-isolation

Training (Mistral 7B):

ACCELERATE_LOG_LEVEL=info accelerate launch \
  --config_file accelerate_configs/deepspeed_zero3.yaml \
  scripts/run_simpo.py \
  training_configs/mistral-7b-base-simpo.yaml

Common workflows

Workflow 1: Train from base model (Mistral 7B)

Config (mistral-7b-base-simpo.yaml):

# Model
model_name_or_path: mistralai/Mistral-7B-v0.1
torch_dtype: bfloat16

# Dataset
dataset_mixer:
  HuggingFaceH4/ultrafeedback_binarized: 1.0
dataset_splits:
  - train_prefs
  - test_prefs

# SimPO hyperparameters
beta: 2.0                  # Reward scaling (2.0-10.0)
gamma_beta_ratio: 0.5       # Target margin (0-1)
loss_type: sigmoid          # sigmoid or hinge
sft_weight: 0.0             # Optional SFT regularization

# Training
learning_rate: 5e-7         # Critical: 3e-7 to 1e-6
num_train_epochs: 1
per_device_train_batch_size: 1
gradient_accumulation_steps: 8

# Output
output_dir: ./outputs/mistral-7b-simpo

Launch training:

accelerate launch --config_file accelerate_configs/deepspeed_zero3.yaml \
  scripts/run_simpo.py training_configs/mistral-7b-base-simpo.yaml

Workflow 2: Fine-tune instruct model (Llama 3 8B)

Config (llama3-8b-instruct-simpo.yaml):

model_name_or_path: meta-llama/Meta-Llama-3-8B-Instruct

dataset_mixer:
  argilla/ultrafeedback-binarized-preferences-cleaned: 1.0

beta: 2.5
gamma_beta_ratio: 0.5
learning_rate: 5e-7
sft_weight: 0.1             # Add SFT loss to preserve capabilities

num_train_epochs: 1
per_device_train_batch_size: 2
gradient_accumulation_steps: 4
output_dir: ./outputs/llama3-8b-simpo

Read the full file on GitHub · 224 lines

Files

What ships with it

3 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 7d ago First seen · 224 lines · 14 tokens per session scan A 4ebaa42679f7

Subscribe to this mod's changes

simpo-training is a skill published in the GitHub repository MilkyWay008/Hermes-OTG (15 stars, last pushed 26d ago), licensed MIT. It adds 14 tokens to every session and 1,631 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. It is 98% identical to simpo, differing in 2 lines, and is treated as a copy.

Related

Other skills, from other repositories

hunt-llm-ai

Hunt LLM/AI feature bugs — prompt injection, indirect injection, exfiltration viatool-use/markdown, ASCII smuggling, agentic AI security (OWASP Agentic Apps 2026, ASI01-ASI10). Patterns: direct injection ('ignore previous instructions'), indirect injection via documents/web pages/email the model reads, ASCII smuggling…

uphiago/recon-skills · 256 tokens

lijigang-skill

A Chinese-language approach to writing precise, highly structured prompts, sometimes using Lisp-like notation. It combines concise wording, philosophical questioning, and a process for defining roles, conditions, output formats, and revisions.

momozi1996/awesome-ai-persona-skills · 169 tokens

baoyu-skill

A Chinese-language approach to explaining AI tools and writing prompts—instructions that tell an AI what you want. It emphasizes step-by-step teaching, hands-on testing, plain-language technical explanations, and organized knowledge sharing.

momozi1996/awesome-ai-persona-skills · 153 tokens

hf-model-card-research

Extract structured metadata — downloads, likes, benchmark claims, file sizes, author statements — from HuggingFace model cards. Used when the user asks you to "check these models on HF", "pull benchmarks for these variants", or "compare what authors claim.".

AtlasOmnia/donna-starter · 35 tokens

llm-wiki

Build and maintain a Karpathy-style LLM knowledge base — a self-compiling Obsidian markdown wiki where an Agent ingests raw sources, compiles cross-linked concept/entity/summary pages, answers queries against the corpus, lints the graph for health, and audits in-context human feedback filed from Obsidian or the local…

JeanBaissari/llm-wiki-monorepo · 185 tokens

interactive-prompt-analyzer

World-class prompt analyzer v3: multi-modal, predictive, self-improving, context-aware, with real-time cost estimation, counterfactual reasoning, cross-session learning, adversarial testing, and autonomous optimization.

sloemo01/hermes-skills-bundle · 49 tokens