Proof pending. Core topic summary fields are still materializing.
Current research on language models is increasingly focused on enhancing accessibility and performance for low-resource languages, as evidenced by recent work that introduces efficient training pipelines for small language models tailored to specific linguistic needs. Innovations like value-aware numerical representations and modular reasoning frameworks are addressing fundamental limitations in numerical understanding and reasoning efficiency, promising to improve model robustness across various tasks. Additionally, the exploration of multimodal capabilities is gaining traction, with new paradigms that integrate visual inputs and advanced search techniques to enhance comprehension and interaction in noisy environments. This shift not only aims to bridge gaps in language representation but also to optimize performance in real-world applications, such as legal document processing and complex question answering. As the field evolves, the emphasis on tailored, efficient models signals a move towards practical deployment in diverse linguistic and contextual settings, potentially transforming how language technology serves global communities.
We present Kakugo, a novel and cost-effective pipeline designed to train general-purpose Small Language Models (SLMs) for low-resource languages using only the language name as input. By using a large...
Reasoning with large language models often benefits from generating multiple chains-of-thought, but existing aggregation strategies are typically trajectory-level (e.g., selecting the best trace or vo...
Transformer-based language models often achieve strong results on mathematical reasoning benchmarks while remaining fragile on basic numerical understanding and arithmetic operations. A central limita...
The dominance of large multilingual foundation models has widened linguistic inequalities in Natural Language Processing (NLP), often leaving low-resource languages underrepresented. This paper introd...
This technical report presents Sabiá-4 and Sabiazinho-4, a new generation of Portuguese language models with a focus on Brazilian Portuguese language. The models were developed through a four-stage tr...
We introduce A.X K1, a 519B-parameter Mixture-of-Experts (MoE) language model trained from scratch. Our design leverages scaling laws to optimize training configurations and vocabulary size under fixe...
Large language models typically represent Chinese characters as discrete index-based tokens, largely ignoring their visual form. For logographic scripts, visual structure carries semantic and phonetic...
Large language models exhibit sycophantic tendencies--validating incorrect user beliefs to appear agreeable. We investigate whether this behavior varies systematically with perceived user demographics...
Masked Diffusion Language Models (MDLMs) promise parallel token generation and arbitrary-order decoding, yet it remains unclear to what extent current models truly realize these capabilities. We chara...
Multimodal large language models (MLLMs) have achieved remarkable success across a broad range of vision tasks. However, constrained by the capacity of their internal world knowledge, prior work has p...
Freshness
Canonical route: /topics
Agent Handoff
Canonical ID language-models | Route /topic/language-models
REST example
curl https://sciencetostartup.com/api/v1/agent-handoff/topic/language-modelsMCP example
{
"tool": "search_papers",
"arguments": {
"query": "Language Models",
"cluster": "Language Models"
}
}source_context
{
"surface": "topic",
"mode": "topic",
"query": "Language Models",
"normalized_query": "language-models",
"route": "/topic/language-models",
"paper_ref": null,
"topic_slug": "language-models",
"benchmark_ref": null,
"dataset_ref": null
}Use This Via API or MCP
Topic pages bundle paper counts, viability trends, author concentration, and top questions into one canonical surface your agents can reference before they open Signal Canvas or create a workspace.