Welcome any related paper

[minghu0830]

✨ In this issue, we welcome any paper that relates to scientific large language models (Sci-LLMs) and data — including pre-training, post-training, evaluation, RAG, and beyond.

👍 We strongly encourage everyone to submit pull requests directly to help reduce our team's maintenance workload.

👉 We look forward to your contribution.

[Masaaki-75]

would like to spotlight: SOSBENCH: Benchmarking Safety Alignment on Scientific Knowledge, a regulation-grounded, hazard-focused benchmark with 3k prompts across six high-risk scientific domains that rigorously evaluates Sci-LLM safety alignment and uncovers persistent vulnerabilities even in state-of-the-art models.

[Opdoop]

would like to add s1: https://huggingface.co/ScienceOne-AI/S1-Base-671B in the Timeline of Sci-LLMs

[Dominic789654]

Perovskite-llm: Knowledge-enhanced large language models for perovskite solar cell research
Accepted by EMNLP 2025 as Findings.
Abstract:

The rapid advancement of perovskite solar cells (PSCs) has led to an exponential growth in research publications, creating an urgent need for efficient knowledge management and reasoning systems in this domain. We present a comprehensive knowledge-enhanced system for PSCs that integrates three key components. First, we develop Perovskite-KG, a domain-specific knowledge graph constructed from 1,517 research papers, containing 23,789 entities and 22,272 relationships. Second, we create two complementary datasets: Perovskite-Chat, comprising 55,101 high-quality question-answer pairs generated through a novel multi-agent framework, and Perovskite-Reasoning, containing 2,217 carefully curated materials science problems. Third, we introduce two specialized large language models: Perovskite-Chat-LLM for domain-specific knowledge assistance and Perovskite-Reasoning-LLM for scientific reasoning tasks. Experimental results demonstrate that our system significantly outperforms existing models in both domain-specific knowledge retrieval and scientific reasoning tasks, providing researchers with effective tools for literature review, experimental design, and complex problem-solving in PSC research.

[qianyi0206]

Perovskite-llm: Knowledge-enhanced large language models for perovskite solar cell research Accepted by EMNLP 2025 as Findings. Abstract:

The rapid advancement of perovskite solar cells (PSCs) has led to an exponential growth in research publications, creating an urgent need for efficient knowledge management and reasoning systems in this domain. We present a comprehensive knowledge-enhanced system for PSCs that integrates three key components. First, we develop Perovskite-KG, a domain-specific knowledge graph constructed from 1,517 research papers, containing 23,789 entities and 22,272 relationships. Second, we create two complementary datasets: Perovskite-Chat, comprising 55,101 high-quality question-answer pairs generated through a novel multi-agent framework, and Perovskite-Reasoning, containing 2,217 carefully curated materials science problems. Third, we introduce two specialized large language models: Perovskite-Chat-LLM for domain-specific knowledge assistance and Perovskite-Reasoning-LLM for scientific reasoning tasks. Experimental results demonstrate that our system significantly outperforms existing models in both domain-specific knowledge retrieval and scientific reasoning tasks, providing researchers with effective tools for literature review, experimental design, and complex problem-solving in PSC research.

This looks like a very impressive and important project. Congratulations on the EMNLP acceptance!

Are there any plans to open-source the code and the Perovskite-KG dataset? I would be very interested in exploring them. Thanks for sharing!

[Dominic789654]

Perovskite-llm: Knowledge-enhanced large language models for perovskite solar cell research Accepted by EMNLP 2025 as Findings. Abstract:

The rapid advancement of perovskite solar cells (PSCs) has led to an exponential growth in research publications, creating an urgent need for efficient knowledge management and reasoning systems in this domain. We present a comprehensive knowledge-enhanced system for PSCs that integrates three key components. First, we develop Perovskite-KG, a domain-specific knowledge graph constructed from 1,517 research papers, containing 23,789 entities and 22,272 relationships. Second, we create two complementary datasets: Perovskite-Chat, comprising 55,101 high-quality question-answer pairs generated through a novel multi-agent framework, and Perovskite-Reasoning, containing 2,217 carefully curated materials science problems. Third, we introduce two specialized large language models: Perovskite-Chat-LLM for domain-specific knowledge assistance and Perovskite-Reasoning-LLM for scientific reasoning tasks. Experimental results demonstrate that our system significantly outperforms existing models in both domain-specific knowledge retrieval and scientific reasoning tasks, providing researchers with effective tools for literature review, experimental design, and complex problem-solving in PSC research.

This looks like a very impressive and important project. Congratulations on the EMNLP acceptance!

Are there any plans to open-source the code and the Perovskite-KG dataset? I would be very interested in exploring them. Thanks for sharing!

Hi, thanks a lot for your interest and the positive feedback!

That's a great question. We're actively working on the next phase of the project, so we don't have plans for a public release just yet. Please stay tuned for future announcements!