Search arXiv⌕ Search
Home/Archive/arXiv

arXiv · 2403.00823

Adapting to Teammates in a Cooperative Language Game

Christopher Archibald·Spencer Brosnahan
View original source ↗Read source document ↗Export metadata ↓

Abstract

The game of Codenames has recently emerged as a domain of interest for intelligent agent design. The game is unique due to the way that language and coordination between teammates play important roles. Previous approaches to designing agents for this game have utilized a single internal language model to determine action choices. This often leads to good performance with some teammates and inferior performance with other teammates, as the agent cannot adapt to any specific teammate. In this paper we present the first adaptive agent for playing Codenames. We adopt an ensemble approach with the goal of determining, during the course of interacting with a specific teammate, which of our internal expert agents, each potentially with its own language model, is the best match. One difficulty faced in this approach is the lack of a single numerical metric that accurately captures the performance of a Codenames team. Prior Codenames research has utilized a handful of different metrics to evaluate agent teams. We propose a novel single metric to evaluate the performance of a Codenames team, whether playing a single team (solitaire) game, or a competitive game against another team. We then present and analyze an ensemble agent which selects an internal expert on each turn in order to maximize this proposed metric. Experimental analysis shows that this ensemble approach adapts to individual teammates and often performs nearly as well as the best internal expert with a teammate. Crucially, this success does not depend on any previous knowledge about the teammates, the ensemble agents, or their compatibility. This research represents an important step to making language-based agents for cooperative language settings like Codenames more adaptable to individual teammates.

Explore related subjects

cs.AIcs.CL

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Christopher Archibald, Spencer Brosnahan. 2024-02-26. Adapting to Teammates in a Cooperative Language Game. https://arxiv.org/abs/2403.00823

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

DeepFeature: LLM-Empowered Context-aware Feature Generation for Wearable Biosignals

Biosignals collected from wearable devices are widely utilized in healthcare applications. Machine learning models used in these applications often rely on features extracted from biosignals due to their effectiveness, lower data dimensionality, and wide compatibility across various model architectures. However, existing feature extraction methods often lack task-specific contextual knowledge, struggle to identify optimal features in high-dimensional combinatorial feature space, and are prone to automated code generation and execution errors. In this paper, we propose DeepFeature, the first LLM-empowered, context-aware feature generation framework for wearable biosignals. DeepFeature introduces a multi-source feature generation mechanism that integrates the inherent ability of LLMs, expert knowledge and inter-feature interactions. It also employs an iterative feature refinement process that uses feature assessment-based feedback for feature re-selection. Additionally, DeepFeature utilizes a robust multi-layer filtering and verification approach for feature description-to-code translation to ensure that the feature extraction functions run without crashing. Experimental evaluation results show that DeepFeature achieves the highest average AUROC across eight tasks under both sample-level and subject-level settings, outperforming the best baselines by 4.60% and 4.61%, respectively. DeepFeature achieves the most pronounced gains on the PPG-BP tasks, while remaining competitive with the best-performing baselines on Epilepsy, WESAD, and our self-collected SEN dataset.

cs.AI↗
arXiv2026

Modality-Guided Mixture of Structured Experts with Entropy-Triggered Routing for Multimodal Recommendation

Multimodal recommenders combine collaborative behavior with visual and textual item evidence, whose usefulness varies across user-item interactions. Independently trained source-specific diagnostic probes partition held-out interactions into behavior-, appearance-, semantics-, and mixed-evidence regimes across five benchmarks, within which capacity-matched fixed fusion rules exhibit systematic regime-dependent performance crossovers. This diagnostic observation motivates MAGNET, a multimodal graph recommender with two core mechanisms. First, a calibrated expert bank organizes trainable experts by anchor source (behavior, appearance, or semantics) and fusion family (dominant, balanced, or complementary), while an interaction-conditioned router selects a sparse composition using all three evidence sources. Second, an entropy-triggered, coverage-aware progressive schedule decouples population-level routing-mass coverage from per-instance decisiveness, transitioning from broad routing exploration to confident specialization once sufficient coverage is sustained, while preserving coverage thereafter. MAGNET separately encodes the observed interaction graph and a filtered content-induced structural view, applying cross-view alignment after independent propagation. We evaluate MAGNET on four Amazon domains and the non-Amazon MicroLens-100K benchmark, including tail-item and low-history warm-start user evaluation alongside matched expert-design controls. Under the shared-feature, fixed-split protocol, MAGNET-DV exceeds the strongest protocol-compatible non-MAGNET baseline in every reported main-table cell, supported by seed-wise difference tests. Both the fixed and KL-anchored structured variants achieve higher five-seed mean NDCG@20 than matched homogeneous and free-mixture alternatives across all five datasets. MAGNET supports route-level diagnostics through explicit expert semantics.

cs.AI↗
arXiv2026

From Refusal Tokens to Refusal Control: Discovering and Steering Category-Specific Refusal Directions

Language models are commonly fine-tuned for safety alignment to refuse harmful prompts. One approach fine-tunes them to generate categorical refusal tokens that distinguish different refusal types before responding. In this work, we leverage a version of Llama 3 8B fine-tuned with these categorical refusal tokens to enable inference-time control over fine-grained refusal behavior, improving both safety and reliability. We show that refusal token fine-tuning induces separable, category-aligned directions in the residual stream, which we extract and use to construct categorical steering vectors with a lightweight probe that determines whether to steer toward or away from refusal during inference. In addition, we introduce a learned low-rank combination that mixes these category directions in a whitened, orthonormal steering basis, resulting in a single controllable intervention under activation-space anisotropy, and show that this intervention is transferable across same-architecture model variants without additional training. Across benchmarks, both categorical steering vectors and the low-rank combination consistently reduce over-refusals on benign prompts while increasing refusal rates on harmful prompts, highlighting their utility for multi-category refusal control.

cs.AI↗