Wenzhe Li
@WenzheLiTHU
Followers
72
Following
33
Media
1
Statuses
15
Ph.D. student @Princeton ECE Reinforcement Learning
Joined August 2021
Excited to share our recent work LLM Economist! Generative agents can be a powerful tool to simulate and empower social decision makingš.
š NewāÆpreprint! .š¤ Can one agent ānudgeā a synthetic civilization of Censusāgrounded agents toward higher social welfareāall by optimizing utilities inācontext? Meet the LLMāÆEconomist ā
0
0
6
RT @kaiqu_liang: š¤ Feel like your AI is bullshitting you? Itās not just you. šØ We quantified machine bullshit š©. Turns out, aligning LLMsā¦.
0
108
0
RT @sethkarten: Can a Large Language Model (LLM) with zero PokĆ©mon-specific training achieve expert-level performance in competitive PokĆ©moā¦.
0
36
0
RT @omarsar0: Rethinking Mixture-of-Agents. This work investigates whether mixing different LLMs is truly beneficial. They also propose Seā¦.
0
77
0
RT @Yong18850571: š Introducing Goedel-Prover: A 7B LLM achieving SOTA open-source performance in automated theorem proving! š„. ā
Improvingā¦.
0
67
0
RT @zhezeng0908: š¢ Iām recruiting PhD students @CS_UVA for Fall 2025!.šÆ Neurosymbolic AI, probabilistic ML, trustworthiness, AI for scienceā¦.
zzeng.me
Assistant Professor
0
73
0
RT @chijinML: Our team's LLM ensemble method ranked first on AlpacaEval 2.0!š joint work with @WenzheLiTHU, Yong Lin, @xiamengzhou. More deā¦.
0
16
0
0
0
1
While it is feasible to build a single-agent AI that is capable of passing through all 15 levels of the Street Fighter II game (as shown in our project page: , All these MARL methods can be exploited easily rather than learning exact NE policies.
sites.google.com
Abstract Recent advances in reinforcement learning (RL) heavily rely on a variety of well-designed benchmarks, which provide environmental platforms and consistent criteria to evaluate existing and...
1
0
2
Wondering how existing MARL algorithms perform in pure competitive settings? We will present FightLadder at ICML 2024, a competitive MARL benchmark based on popular fighting games, to evaluate them from a game-theoretic perspective:
2
6
10
RT @chijinML: Ever wonder how to play multiplayer games (>2 players, such as Mahjong, Poker) well and what would be the ultimate solution?ā¦.
0
11
0