Multilingual GSM-Symbolic: What determines capability transfer across languages?
Paper • 2610.03367 • Published • 50
Sailor2
community
AI & ML interests
Open language models for South-East Asia
Recent Activity
View all activity
jjzha
authored a
paper 2 days ago
afaji
submitted a
paper to Daily Papers 3 days ago
patrickamadeus
authored 3
papers 4 days ago
patrickamadeus
submitted a
paper to Daily Papers 5 days ago
ThuraAung1601
authored 3
papers about 1 month ago
Last Translation Benchmark
Paper • 2609.04173 • Published • 46
SEA-BED: Southeast Asia Embedding Benchmark
Paper • 2508.12243 • Published • 1
myNER: Contextualized Burmese Named Entity Recognition with Bidirectional LSTM and fastText Embeddings via Joint Training with POS Tagging
Paper • 2504.04038 • Published
jjzha
authored a
paper about 1 month ago
Muennighoff
submitted a
paper to Daily Papers about 1 month ago
Muennighoff
submitted a
paper to Daily Papers about 2 months ago
dreamerdeo
authored a
paper about 2 months ago
Post
483
Shipped v0.1.2 of vtx — a minimalist coding agent for the terminal.
Most agentic CLIs ship 10k+ token system prompts. Vtx is ~2,200. Less prompt overhead means more room for your code in the model's context window.
Vtx is a from-scratch Python implementation of the design philosophy behind pi-mono — same principles, pure Python, no transpiled runtime.
What ships out of the box:
→ Textual TUI + headless CLI (vtx -p "fix the failing test")
→ 49 LLM provider gateways, all declared in a single provider.yaml
→ 5 core tools (read / edit / write / bash / find) plus web search and fetch
→ Session tree with compaction, handoff, and resume
→ AGENTS.md / CLAUDE.md auto-discovery
→ Skills system — drop SKILL.md files in .agents/skills/ and they become slash commands
→ Two OAuth flows (GitHub Copilot device flow, OpenAI Codex PKCE)
→ Two-mode permissions: prompt (default) or auto, with a safe-command allowlist
This release adds a proper extension system. Register new LLM-callable tools, intercept tool calls, hook lifecycle events, and add slash commands from a single register(api) function in a Python file under ~/.vtx/agent/extensions/. Extensions can override built-in tools by name and chain handler logic across subscribers.
Apache 2.0. uv tool install vtx-coding-agent and you're running.
GitHub: https://github.com/OEvortex/vtx-coding-agent
PyPI: https://pypi.org/project/vtx-coding-agent
Built in the open. Feedback, extensions, and PRs welcome.
Most agentic CLIs ship 10k+ token system prompts. Vtx is ~2,200. Less prompt overhead means more room for your code in the model's context window.
Vtx is a from-scratch Python implementation of the design philosophy behind pi-mono — same principles, pure Python, no transpiled runtime.
What ships out of the box:
→ Textual TUI + headless CLI (vtx -p "fix the failing test")
→ 49 LLM provider gateways, all declared in a single provider.yaml
→ 5 core tools (read / edit / write / bash / find) plus web search and fetch
→ Session tree with compaction, handoff, and resume
→ AGENTS.md / CLAUDE.md auto-discovery
→ Skills system — drop SKILL.md files in .agents/skills/ and they become slash commands
→ Two OAuth flows (GitHub Copilot device flow, OpenAI Codex PKCE)
→ Two-mode permissions: prompt (default) or auto, with a safe-command allowlist
This release adds a proper extension system. Register new LLM-callable tools, intercept tool calls, hook lifecycle events, and add slash commands from a single register(api) function in a Python file under ~/.vtx/agent/extensions/. Extensions can override built-in tools by name and chain handler logic across subscribers.
Apache 2.0. uv tool install vtx-coding-agent and you're running.
GitHub: https://github.com/OEvortex/vtx-coding-agent
PyPI: https://pypi.org/project/vtx-coding-agent
Built in the open. Feedback, extensions, and PRs welcome.
jjzha
authored 4
papers 4 months ago
CommonLID: Re-evaluating State-of-the-Art Language Identification Performance on Web Data
Paper • 2601.18026 • Published
UniSkill: A Dataset for Matching University Curricula to Professional Competencies
Paper • 2603.03134 • Published
WorkRB: A Community-Driven Evaluation Framework for AI in the Work Domain
Paper • 2604.13055 • Published
CroCo: Cross-Lingual Contrastive Preference Tuning on Self-Generations
Paper • 2605.26293 • Published • 3
jjzha
submitted a
paper to Daily Papers 4 months ago
jojo-ai-mst
posted an update 6 months ago
Post
221
🇲🇲 Releasing the Myanmar Tuberculosis Instruction Dataset — a Myanmar–English parallel corpus for medical NLP in one of the lowest-resourced language settings in Southeast Asia.
Most TB datasets are either structured clinical data or English-only research corpora. This one fills a different gap: instructional, guideline-based content in Burmese, formatted for instruction tuning and medical QA.
### What's inside
- 2,043 instruction–response pairs
- Myanmar–English parallel
- 7 TB domains: treatment, diagnostics, drug management, MDR-TB, infection control, patient education, healthcare worker training
- Sourced from WHO guidelines, Myanmar NTP protocols, and standard medical references
- MIT licensed
Useful for
- Fine-tuning Myanmar-language medical LLMs
- TB question answering
- Translation evaluation in a medical domain
- General low-resource medical NLP
👉 jojo-ai-mst/Myanmar-Tuberculosis-Guidelines-Instructions
Built by Min Si Thu and Khin Myat Noe. Feedback welcome — especially from anyone working on SEA medical AI or Burmese NLP.
#MedicalAI #LowResourceNLP #Myanmar #Burmese #Tuberculosis #InstructionTuning
Most TB datasets are either structured clinical data or English-only research corpora. This one fills a different gap: instructional, guideline-based content in Burmese, formatted for instruction tuning and medical QA.
### What's inside
- 2,043 instruction–response pairs
- Myanmar–English parallel
- 7 TB domains: treatment, diagnostics, drug management, MDR-TB, infection control, patient education, healthcare worker training
- Sourced from WHO guidelines, Myanmar NTP protocols, and standard medical references
- MIT licensed
Useful for
- Fine-tuning Myanmar-language medical LLMs
- TB question answering
- Translation evaluation in a medical domain
- General low-resource medical NLP
from datasets import load_dataset
ds = load_dataset("jojo-ai-mst/Myanmar-Tuberculosis-Guidelines-Instructions")👉 jojo-ai-mst/Myanmar-Tuberculosis-Guidelines-Instructions
Built by Min Si Thu and Khin Myat Noe. Feedback welcome — especially from anyone working on SEA medical AI or Burmese NLP.
#MedicalAI #LowResourceNLP #Myanmar #Burmese #Tuberculosis #InstructionTuning