Possible Stylesheet or format spillage

arXiv / html_feedback

Supports a student project developing a UI for feedback on arXiv articles rendered as html.

MIT License

18 stars 3 forks source link

Description

table or figure not rendered properly. Paper is https://arxiv.org/html/2404.01230v1 section 3

(Optional:) Please add any files, screenshots, or other information here.

3 Scenarios: Where to Apply Strategic Reasoning with LLMs \forestset

every leaf node/.style= if n children=0#1 , every tree node/.style= if n children=0minimum width=1em#1 , {forest} for tree= every leaf node=my leaf, font=, every tree node=my node, font=, l sep-=4.5pt, l-=1.pt, anchor=west, inner sep=2pt, minimum height = 15pt, font=, l sep=8pt, s sep=2pt, fit=tight, grow’=east, edge=ultra thin, parent anchor=east, child anchor=west, if n children=0tier=last, edge path= [draw, \forestoptionedge] (!u.parent anchor) – +(5pt,0) |- (.child anchor)\forestoptionedge label; , if=isodd(n_children()) for children= if=equal(n,(n_children("!u")+1)/2)calign with current [Scenarios, draw=gray, color=gray!100, fill=gray!15, very thick, text=black, text width=1.4cm, minimum height = 18pt [ Societal Simulation, color=brightlavender!100, fill=brightlavender!15, very thick, text=black [ Societal behavior, color=brightlavender!100, fill=brightlavender!15, very thick, text=black [ BigToM(Gandhi et al., 2024), SOTOPIA (Zhou et al., 2023), UGI (Xu et al., 2023a), Suzuki & Arita (2024), OpenToM(Xu et al., 2024a) , color=brightlavender!100, very thick, text=black, tier=E ] ] [ Debate & Negotiation, color=brightlavender!100, fill=brightlavender!15, very thick, text=black [ Fu et al. (2023), Abdelnabi et al. (2023), TRIP (Zhang et al., 2024a), WarAgent (Hua et al., 2023), Flamino et al. (2024), Taubenfeld et al. (2024), Tang et al. (2023), Gemp et al. (2024), Hua et al. (2024), Schneider et al. (2023), Lamparth et al. (2024) , color=brightlavender!100, very thick, text=black, tier=E ] ] ] [ Economic Simulation, color=lightcoral!100, fill=lightcoral!15, very thick, text=black [ Economics, color=lightcoral!100, fill=lightcoral!15, very thick, text=black [ Horton (2023), Xie et al. (2023), Chen et al. (2023b), Li et al. (2023b) , color=lightcoral!100, very thick, text=black, tier=latex forest tier=D ] ] [ Business, color=lightcoral!100, fill=lightcoral!15, very thick, text=black [ Han et al. (2023), TradingGPT (Li et al., 2023c), CompeteAI (Zhao et al., 2023), Chen et al. (2023a), OG-Narrator (Xia et al., 2024) , color=lightcoral!100, very thick, text=black, tier=E ] ] ] [ Game Theory, color=cyan!100, fill=cyan!15, very thick, text=black [ Matrix Game, color=cyan!100, fill=cyan!15, very thick, text=black [ Guo (2023), Phelps & Russell (2023), MAgIC (Xu et al., 2023b), Gandhi et al. (2023), Brookins & DeBacker (2023), Fan et al. (2024) , color=cyan!100, very thick, text=black, tier=E ] ] [ Repeated Game, color=cyan!100, fill=cyan!15, very thick, text=black [ Akata et al. (2023), Alympics (Mao et al., 2023), K-Level Reasoning (Zhang et al., 2024c), Wu et al. (2024b), γ-Bench (Huang et al., 2024b) , color=cyan!100, very thick, text=black, tier=E ] ] ] [ Gaming, color=lightgreen!100, fill=lightgreen!15, very thick, text=black [ Conversational Games, color=lightgreen!100, fill=lightgreen!15, very thick, text=black [ WereWolf, color=lightgreen!100, fill=lightgreen!15, very thick, text=black, text width=1.2cm [ Xu et al. (2023d), Xu et al. (2023c), Thinker (Wu et al., 2024a) , color=lightgreen!100, very thick, text=black, tier=D, text width=4.45cm] ] [ Avalon, color=lightgreen!100, fill=lightgreen!15, very thick, text=black, text width=1.2cm [ Avalonbench Light et al. (2023), ReCon (Wang et al., 2023a), (Lan et al., 2023) , color=lightgreen!100, very thick, text=black, tier=D, text width=4.45cm] ] [ Diplomacy, color=lightgreen!100, fill=lightgreen!15, very thick, text=black, text width=1.2cm [ Welfare diplomacy (Mukobi et al., 2023), Cicero († et al.(2022)(FAIR)†, Bakhtin, Brown, Dinan, Farina, Flaherty, Fried, Goff, Gray, Hu, et al., FAIR) , color=lightgreen!100, very thick, text=black, tier=D, text width=4.45cm] ] [ Other, color=lightgreen!100, fill=lightgreen!15, very thick, text=black, text width=1.2cm [ MAgIC(Xu et al., 2023b), Hoodwinked (O’Gara, 2023), Tsai et al. (2023), Li et al. (2023a) , color=lightgreen!100, very thick, text=black, tier=D, text width=4.45cm] ] ] [ Board & Card Game, color=lightgreen!100, fill=lightgreen!15, very thick, text=black [ Poker, color=lightgreen!100, fill=lightgreen!15, very thick, text=black, text width=1.2cm [ SuspicionAgen (Guo et al., 2023), PokerGPT (Huang et al., 2024a), Agent-Pro (Zhang et al., 2024b) , color=lightgreen!100, very thick, text=black, tier=D, text width=4.45cm] ] [ Chess, color=lightgreen!100, fill=lightgreen!15, very thick, text=black, text width=1.2cm [ ChessGPT (Feng et al., 2024), Kuo et al. (2023), Feng et al. (2023), Guo et al. (2024a) , color=lightgreen!100, very thick, text=black, tier=D, text width=4.45cm] ] ] [ Electronic Game, color=lightgreen!100, fill=lightgreen!15, very thick, text=black [ LLM-Co (Agashe et al., 2023), TextStarCraft II (Ma et al., 2023), SwarmBrain (Shao et al., 2024), PokéLLMon (Hu et al., 2024), CivRealm (Qi et al., 2024), DOOM(de Wynter, 2024) , color=lightgreen!100, very thick, text=black ] ] ] ] Figure 2: Taxonomy of scenarios of strategic reasoning by LLM-based agents.

(Required) What is this issue most closely related to? Select one.

Choose One

Internal issue ID

b3022970-8f23-4933-93a5-6bb9fe78f6c5

Paper URL

https://arxiv.org/html/2404.01230v1

Browser

Firefox/122.0

Device Type

Desktop

Location in document: S3.F2.2

Selected HTML: \forestset

every leaf node/.style= if n children=0#1 , every tree node/.style= if n children=0minimum width=1em#1 , {forest} for tree= every leaf node=my leaf, font=, every tree node=my node, font=, l sep-=4.5pt, l-=1.pt, anchor=west, inner sep=2pt, minimum height = 15pt, font=, l sep=8pt, s sep=2pt, fit=tight, grow’=east, edge=ultra thin, parent anchor=east, child anchor=west, if n children=0tier=last, edge path= [draw, \forestoptionedge] (!u.parent anchor) – +(5pt,0) |- (.child anchor)\forestoptionedge label; , if=isodd(n_children()) for children= if=equal(n,(n_children("!u")+1)/2)calign with current [Scenarios, draw=gray, color=gray!100, fill=gray!15, very thick, text=black, text width=1.4cm, minimum height = 18pt [ Societal Simulation, color=brightlavender!100, fill=brightlavender!15, very thick, text=black [ Societal behavior, color=brightlavender!100, fill=brightlavender!15, very thick, text=black [ BigToM(Gandhi et al., 2024), SOTOPIA (Zhou et al., 2023), UGI (Xu et al., 2023a), Suzuki & Arita (2024), OpenToM(Xu et al., 2024a) , color=brightlavender!100, very thick, text=black, tier=E ] ] [ Debate &
Negotiation, color=brightlavender!100, fill=brightlavender!15, very thick, text=black [ Fu et al. (2023), Abdelnabi et al. (2023), TRIP (Zhang et al., 2024a), WarAgent (Hua et al., 2023), Flamino et al. (2024), Taubenfeld et al. (2024), Tang et al. (2023), Gemp et al. (2024), Hua et al. (2024), Schneider et al. (2023), Lamparth et al. (2024) , color=brightlavender!100, very thick, text=black, tier=E ] ] ] [ Economic Simulation, color=lightcoral!100, fill=lightcoral!15, very thick, text=black [ Economics, color=lightcoral!100, fill=lightcoral!15, very thick, text=black [ Horton (2023), Xie et al. (2023), Chen et al. (2023b), Li et al. (2023b) , color=lightcoral!100, very thick, text=black, tier=latex forest tier=D ] ] [ Business, color=lightcoral!100, fill=lightcoral!15, very thick, text=black [ Han et al. (2023), TradingGPT (Li et al., 2023c), CompeteAI (Zhao et al., 2023), Chen et al. (2023a), OG-Narrator (Xia et al., 2024) , color=lightcoral!100, very thick, text=black, tier=E ] ] ] [ Game Theory, color=cyan!100, fill=cyan!15, very thick, text=black [ Matrix Game, color=cyan!100, fill=cyan!15, very thick, text=black [ Guo (2023), Phelps & Russell (2023), MAgIC (Xu et al., 2023b), Gandhi et al. (2023), Brookins & DeBacker (2023), Fan et al. (2024) , color=cyan!100, very thick, text=black, tier=E ] ] [ Repeated Game, color=cyan!100, fill=cyan!15, very thick, text=black [ Akata et al. (2023), Alympics (Mao et al., 2023), K-Level Reasoning (Zhang et al., 2024c), Wu et al. (2024b), $\gamma$ -Bench (Huang et al., 2024b) , color=cyan!100, very thick, text=black, tier=E ] ] ] [ Gaming, color=lightgreen!100, fill=lightgreen!15, very thick, text=black [ Conversational Games, color=lightgreen!100, fill=lightgreen!15, very thick, text=black [ WereWolf, color=lightgreen!100, fill=lightgreen!15, very thick, text=black, text width=1.2cm [ Xu et al. (2023d), Xu et al. (2023c), Thinker (Wu et al., 2024a) , color=lightgreen!100, very thick, text=black, tier=D, text width=4.45cm] ] [ Avalon, color=lightgreen!100, fill=lightgreen!15, very thick, text=black, text width=1.2cm [ Avalonbench Light et al. (2023), ReCon (Wang et al., 2023a), (Lan et al., 2023) , color=lightgreen!100, very thick, text=black, tier=D, text width=4.45cm] ] [ Diplomacy, color=lightgreen!100, fill=lightgreen!15, very thick, text=black, text width=1.2cm [ Welfare diplomacy (Mukobi et al., 2023), Cicero († et al.(2022)(FAIR)†, Bakhtin, Brown, Dinan, Farina, Flaherty, Fried, Goff, Gray, Hu, et al., FAIR) , color=lightgreen!100, very thick, text=black, tier=D, text width=4.45cm] ] [ Other, color=lightgreen!100, fill=lightgreen!15, very thick, text=black, text width=1.2cm [ MAgIC(Xu et al., 2023b), Hoodwinked (O’Gara, 2023), Tsai et al. (2023), Li et al. (2023a) , color=lightgreen!100, very thick, text=black, tier=D, text width=4.45cm] ] ] [ Board & Card Game, color=lightgreen!100, fill=lightgreen!15, very thick, text=black [ Poker, color=lightgreen!100, fill=lightgreen!15, very thick, text=black, text width=1.2cm [ SuspicionAgen (Guo et al., 2023), PokerGPT (Huang et al., 2024a), Agent-Pro (Zhang et al., 2024b) , color=lightgreen!100, very thick, text=black, tier=D, text width=4.45cm] ] [ Chess, color=lightgreen!100, fill=lightgreen!15, very thick, text=black, text width=1.2cm [ ChessGPT (Feng et al., 2024), Kuo et al. (2023), Feng et al. (2023), Guo et al. (2024a) , color=lightgreen!100, very thick, text=black, tier=D, text width=4.45cm] ] ] [ Electronic Game, color=lightgreen!100, fill=lightgreen!15, very thick, text=black [ LLM-Co (Agashe et al., 2023), TextStarCraft II (Ma et al., 2023), SwarmBrain (Shao et al., 2024), PokéLLMon (Hu et al., 2024), CivRealm (Qi et al., 2024), DOOM(de Wynter, 2024) , color=lightgreen!100, very thick, text=black ] ] ] ]

arXiv / html_feedback