#livecodebench - Tagar Twitter

KEYLOCK @keylockcom

sekitar 2 bulan yang lalu

https://t.co/OZEY2RAyvo #code #ai #livecode #livecodebench #domainforsale

0

2

1

0

38

KEYLOCK @keylockcom

sekitar 2 bulan yang lalu

https://t.co/OZEY2RAyvo #code #ai #livecode #livecodebench #domainforsale

0

2

1

0

75

KEYLOCK @keylockcom

2 bulan yang lalu

https://t.co/Ii5u45hySC #code #ai #livecode #livecodebench #domain

0

1

0

79

Leafy🍃 @Leafy_Fi

12 bulan yang lalu

#LiveCodeBench Pro is doing something smart that more benchmarks should copy: pulls problems in real-time from actual competitions (Codeforces, ICPC, IOI) Olympic medalists explain where/why models break avoids the whole data pollution mess turns out there's still a massive gap between SOTA models and top human reasoners 👀 @SentientAGI

Leafy_Fi's tweet photo. #LiveCodeBench Pro is doing something smart that more benchmarks should copy:

pulls problems in real-time from actual competitions (Codeforces, ICPC, IOI)

Olympic medalists explain where/why models break

avoids the whole data pollution mess

turns out there's still a massive gap between SOTA models and top human reasoners 👀
@SentientAGI

1

3

1

0

72

its_iVincent @Vincent_The2nd

sekitar 1 tahun yang lalu

Bullish on their mission for open AGI not controlled by big corps... closed Source Check the full deets: - Leaderboard & Docs: https://t.co/JDN4ZUb4EC - Paper: https://t.co/FW8iUPgcAI - Toolkit/GitHub: https://t.co/5dCyziGmBv #SentientAGI #LiveCodeBench #NeurIPS #AI #OpenAGI

0

4

0

53

its_iVincent @Vincent_The2nd

sekitar 1 tahun yang lalu

#SentientAGI #LiveCodeBench #NeurIPS #AI #OpenAGI The Problem with Current LLM Coding Benchmarks: Most evals are too forgiving leaked problems, easy prompts, no real limits. LLMs score high but flop in the real world.

1

2

0

92

武智倫太郎 @aiethics496

lebih dari 1 tahun yang lalu

AIの中身を問わない社会は、知性ではなく演出に支配される。孫正義・https://t.co/Pz7vJrrBUf・攻殻機動隊を通じて構造を暴きました。 #AI神話解体 #孫正義批評 ##生成AI詐欺 #BuilderAI #攻殻機動隊 #脱知性社�� #LiveCodeBench https://t.co/Ki4HsWtlcN https://t.co/X0vVqf9Zn2

0

155

15

0

2K

AniWisp - Gaming and Creative Platform @aniwispcom

lebih dari 1 tahun yang lalu

#Alibaba Releases #QwQ‑Max‑Preview, a Reasoning Model Based on #Qwen 2.5‑Max. Beats #DeepSeek‑R1 & #OpenAI's #o1 in #LiveCodeBench ⸺ Solves Complex Mathematical & Logical Problems ⸺ Enables Thinking Mode in Qwen Chat ⸺ Can Combine Different Data Types https://t.co/alpCogLmpm

0

2

0

204

Sean Hughes @hughesthe1st

lebih dari 2 tahun yang lalu

StarCoder2 continues to punch above its weight, with a strong performance relative to much larger open base models on the new #LiveCodeBench leaderboard.

Naman Jain

@StringChaos

lebih dari 2 tahun yang lalu

🌟 OSS Coding Models for LCB 🏆 DeepSeek (33B), StarCoder2 (15B), and CodeLLaMa (34B) emerge as the top base models 💫 Finetuning: 👩‍💻 Boosts both LCB & HumanEval performance ⚠️ May overfit to HumanEval-style problems ➡️ Need to diversify open fine-tuning data for robust gains

1

13

1

2K

0

11

2

0

874

Tianjun Zhang

@tianjun_zhang

lebih dari 2 tahun yang lalu

Congrats to @Replit on the release!! Super glad to see that the concept of lively updated coding benchmarks proposed in #LiveCodeBench has been adopted in their design criteria🚀 Checkout more details about LiveCodeBench at https://t.co/XLqkXoygn0🫡