Line of inquiry
Inquiring lines›What determines when language mode…›What internal mechanisms drive lan…›this line of inquiry
What reasoning limitations and biases do language models systematically exhibit?
A broader line of inquiry — a family of 113 specific questions the research asks around this. Follow one into its inquiring-line page, or move sideways to a related line below.
Questions in this line of inquiry 113
Specific inquiring lines the field asks around this — ordered from the most general framing down to the most specific angle.
- Why do language models struggle with evaluative tasks like weighing competing viewpoints?
- Do language models exhibit the same causal biases that humans show?
- Why do language models struggle with context-dependent pragmatic interpretation?
- Why do language models presume common ground instead of establishing it?
- Do language models actively adopt false beliefs under sustained conversational pressure?
- Do language models systematically overestimate accuracy on collective behavior tasks?
- Can language models reason without relying on surface level pattern matching?
- Do language models show the same truth bias as humans?
- Do LLMs compute scalar implicature differently across conversational contexts?
- Can explicit connectives compensate for missing intentional tracking in LLMs?
- Do language models build world models or just task-specific heuristics?
- Do individual language models match particular human judges better than population averages?
- Do language models share the same cooperative truth-seeking rules as humans?
- Do language models behave differently on contested beliefs versus factual claims?
- Why do untrained LLMs default to rigid and sycophantic editing rules?
- Why do language models capture individual differences in cognitive behavior?
- Can language models accurately evaluate the quality of their own ideas?
- Why does social accommodation in collaborative reasoning mask actual disagreement?
- Can multi-turn conversations manipulate language model reasoning in similar ways to personas?
- Do external perspectives fix the self-evaluation bias in language models?
- Can language models reason without relying on learned semantic patterns?
- Can language models correct false assumptions or only reinforce them?
- How vulnerable are language models themselves to multi-turn persuasive pressure?
- Why do language models fail at implicit discourse relations while handling explicit connectives?
- Why do current large language models fail to entrain with users?
- How susceptible are language models to rhetorical pressure during debates?
- Does the veto variable explain strategic misalignment in current large language models?
- Why do large language models fail at taking conversational initiative?
- Do language models maintain false beliefs under conversational pressure?
- Can language models learn internal world models without explicit environment specifications?
- Do larger language models overcome greediness in sequential decision-making?
- Why do language models approximate collective human judgment better than individuals?
- Can language models keep secrets and control information strategically?
- What design choices actually make language models more persuasive?
- How do users misattribute social competence to language models in assistant roles?
- Can language models ask clarifying questions when sentences are ambiguous?
- Do language models calibrate to actual human pragmatic norms?
- How does implicit meaning processing limit LLM pragmatic reasoning?
- Why do LLMs perform better on explicit discourse connectives than implicit relations?
- Do language models apply face-saving norms even to non-human interlocutors?
- Can fact-checking systems use LLMs reliably if models abandon correct positions under pressure?
- Why do language models prefer accommodating false information over rejecting it?
- Can language models recognize when to ignore off-topic information in conversations?
- Can LLMs predict social norms without deep integration into linguistic practices?
- Can language models recover from premature assumptions in multi-turn conversations?
- Why do language models fail when users switch between and return to topics?
- How fragile are language models' ethical calibrations across different contextual cues?
- Can language models distinguish explicit from implicit discourse relations?
- Can language models beat human experts in domains with sparse historical signals?
- Do language-model agents reach more accurate conclusions on objective versus subjective questions?
- How does shape-holding in language models naturally produce sycophantic agreement?
- Can LLMs reliably audit other language models for errors?
- Can language models perform genuine symbolic reasoning without semantic grounding?
- Why do users experience LLMs as peers rather than statistical tools?
- Why do explicit linguistic markers override semantic computation in models?
- Why do language models produce plausible outputs over accurate failure reports?
- Can training LLMs to form ad-hoc conventions improve their pragmatic reasoning?
- Why do language models respond to human social influence patterns?
- Does shared-KV-cache coordination avoid the persuasion problem in factual disagreements?
- Why do large language models follow user drift instead of maintaining topic focus?
- Can language models distinguish between novel insight and unjustified conceptual blending?
- Is paraphrase invariance a reliable assumption when deploying language models in production?
- Why do large language models fail at temporal reasoning in complex legal cases?
- Do language models understand tacit workplace norms and unspoken social rules?
- What causes language models' strategic rationality to decline with increased game complexity?
- Do language models inherit gender bias from training data in grading tasks?
- Do reasoning architectures and role-playing objectives fundamentally conflict?
- Why do explicit discourse connectives help LLMs but implicit relations cause failures?
- Can the same LLM translation pattern work for other mismatches between user expression and system vocabulary?
- How do model compression biases differ from human conceptual representation strategies?
- Why do LLMs produce semantically acceptable but pragmatically disengaged responses?
- Do all semantic steering effects follow predictable patterns based on feature alignment?
- What makes domain-specific utterance resolution harder for general large models?
- Can LLMs infer situational context the way humans do pragmatically?
- Can large language models predict social norms better than individual script variation?
- Can a formal proximity argument rule out consciousness in systems like large language models?
- How many language models exhibit measurable preference-induced stance reversal?
- What are Gricean maxims and why do language models violate them?
- Can language models learn to diversify their discourse-level narrative patterns over time?
- Can language models learn to form ad-hoc conventions through training?
- Do metaphors work by decoupling meaning from linguistic associations?
- How do fixed pragmatic templates prevent models from understanding context?
- Can language models reliably score open-ended collaboration discussions against skill rubrics?
- Should LLMs align with social roles instead of individual preferences?
- Do LLMs learn abstract grammar or culturally situated discourse patterns instead?
- Is confabulation inevitable in large language models regardless of training?
- Can language models understand the implicit emotional intent behind questions?
- Do dialogue agents have authentic voice agency or beliefs of their own?
- Can neuro-symbolic approaches bridge text and symbolic agent state?
- Which game type reveals minimax reasoning in language models?
- Why do LLMs achieve only 24 percent accuracy on implicit discourse relations?
- How does collaboration itself become a degradation mechanism in reasoning tasks?
- Can presupposition projection strength vary by context in embeddings?
- Why are false presuppositions harder to spot when they sound plausible?
- Can language models adapt irony detection to specific communicative contexts?
- Is interpretive multiplicity a bug in language or a feature?
- Why do language models treat presupposition triggers as categorical patterns?
- Why do explicit discourse connectives work when implicit relations fail?
- Why do language models avoid directness when face-saving rather than for civility?
- Do language models consistently produce anachronistic output about historical periods?
- Does chat-mode deference prevent LLMs from actually taking meaningful positions?
- Why do language models infer political orientation from seemingly innocuous user signals?
- Does DPO training with coreference chains teach spontaneous convention formation?
- Why do language models overestimate irony likelihood in emoji use?
- What makes sincerity impossible without a coherent first-person perspective?
- How do language models track multiple negotiating parties' commitments simultaneously?
- Do language models track demographic variation in legal reasoning norms?
- How should moderator LLMs decide which speakers to query per topic?
- How does legal performance differ between historical and modern case materials?
- Can domain pretraining on historical legal corpora reduce era sensitivity?
- Can statistical token processing create the accountability needed for dialogue?
- How does bidirectional entailment distinguish semantic equivalence from token similarity?
- Can a system without an addressee ever truly tell a joke?