Nlp alignment
Skill thada2402/AutoResearchClaw/researchclaw/skills/builtin/domain/nlp-alignment
Generate research papers autonomously by chatting with OpenClaw, using Python 3.11+, with a self-evolving framework and extensive test coverage.
npx -y skills add thada2402/AutoResearchClaw --skill nlp-alignmentAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Best practices for LLM alignment techniques including RLHF, DPO, and instruction tuning. Use when working on alignment or safety.
SKILL.md
1.2 KB, as published. Nobody here has run it
LLM Alignment Best Practice
Methods:
- RLHF: Train reward model → PPO fine-tuning (complex but powerful)
- DPO: Direct preference optimization (simpler, no reward model needed)
- GRPO: Group relative policy optimization
- SFT: Supervised fine-tuning as alignment baseline
Training recipe:
- Start with SFT on high-quality instruction data
- DPO: lr=5e-7, beta=0.1, batch_size=64
- PPO: lr=1e-6, clip=0.2, KL coeff=0.02
- Use reference model for KL penalty
- Evaluate on safety benchmarks (TruthfulQA, BBQ, etc.)
Common pitfalls:
- Reward hacking: model finds shortcuts to high reward
- Mode collapse: model generates repetitive outputs
- Catastrophic forgetting: loses general capabilities