RubricEM: Meta-RL with Rubric-guided Policy Decomposition beyond Verifiable Rewards
Paper • 2605.10899 • Published • 79
Sk LSSG
REPOCHEAKER
·
AI & ML interests
None yet
Recent Activity
reacted to pankajpandey-dev's post with 👀 about 1 month ago
🇮🇳 Qwen3.5-9B Hindi Instruct — it stops thinking in English
Ask base Qwen3.5-9B a question in Hindi and it burns hundreds of tokens thinking in English inside its think block before a single Devanagari word appears — then code-switches in the answer. I fine-tuned it to close the think block instantly and reply in pure, native Hindi.
✅ Model (16-bit): https://huggingface.co/pankajpandey-dev/qwen3.5-9b-hindi-instruct
✅ GGUF (Q4/Q5/Q8): https://huggingface.co/pankajpandey-dev/qwen3.5-9b-hindi-instruct-GGUF
✅ Try it in the browser: https://huggingface.co/spaces/pankajpandey-dev/qwen3.5-9b-hindi-demo
Recipe: Unsloth + LoRA (r=16, response-only loss) on 12.9k Hindi pairs — AI4Bharat anudesh + dolly-hi + wikiHow-hi + Aya Hindi (human-written). The Q4_K_M is 5.4 GB and runs on a plain laptop CPU.
New in this run vs my earlier models: mixed in long-form native sources (wikiHow) after my last eval showed the fine-tune traded detail for conciseness — this one keeps answers detailed and native.
Part of my weekly 🇮🇳 Hindi LLM Series. Feedback welcome 🙏
#Hindi #IndicNLP #Qwen #GGUF #LocalLLM #Unsloth upvoted a paper 3 months ago
RubricEM: Meta-RL with Rubric-guided Policy Decomposition beyond Verifiable Rewards upvoted a paper 3 months ago
LiSA: Lifelong Safety Adaptation via Conservative Policy InductionOrganizations
upvoted 2
papers 3 months ago