Overbid frequency, self-bidding rate, bankrupt-initiation patterns, and context-dependent offer calibration are failure modes invisible to both static evaluations and aggregate rankings like Elo

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

• Do these failure modes (overbidding, self-bidding, bankrupt initiation) generalise to other economic settings?question0.849
  Remains untested whether the specific LLM failures observed in CATTLE TRADE extend beyond this game.
• Overbidding, self-bidding spirals, and undisciplined bluffing characterise failure.claim0.839
  Concrete failure signatures extracted from traces.
• Behavioural traces surface recurring LLM failure modes including overbidding, self-bidding, bankrupt TC initiation, and weak opponent-state adaptation that never appear in code agents.claim0.835
  LLMs exhibit systematic errors that deterministic logic avoids.
• Does a high self-bidding rate reflect a failure to detect non-competitive contexts or a deliberate escalation?question0.818
  Ambiguity in interpreting the self-bidding metric: from a single trace, cannot distinguish error from aggressive strategy.
• Two heuristic code agents outperform most tested LLMs, and behavioural traces surface recurring LLM failure modes including overbidding, self-bidding, bankrupt TC initiation, and weak opponent-state adaptation.quote0.814
  Abstract sentence summarising performance and failures.
• The structured game logs make failure modes directly observable and quantifiableclaim0.762
  design claim about transparency
• G3-F conditions TC offers on opponent wealth and game context, e.g., 0-value bluffs against bankrupt opponentsfinding0.758
  sophisticated bluff calibration
• Scaling Laws for Activation Steering with Llama 2 Models and Refusal Mechanisms (Ali et al., 2025)concept0.758
  Related work finding larger models more resistant to steering, potentially consistent with ESR in 70B