{"arxiv_id":"2608.12372","incoming":{"papers":0,"sentences":0,"groups":[],"unlabelled_sentences":0},"outgoing":{"papers":19,"sentences":24,"groups":[{"kind":"supports_this_paper","label":"Cites as support","papers":6,"sentences":9,"quotes":[{"arxiv_id":"2507.11473","title":"Chain of Thought Monitorability: A New and Fragile Opportunity for AI Safety","url":"/paper/2507.11473","section":"Limitations of Current Alignment Methods","sentence":"There is still uncertainty about whether the faithfulness of their reasoning chains can ever be truly assessed [ref] , but [ref] empirically show that LLM chain-of-thought rationales “are frequently unfaithful, diverging from the true hidden computations that drive [LLM output]”","intent_p":0.99,"contests_p":null},{"arxiv_id":"2508.06950","title":"Exploring the Limits of Large Language Models as a Reflection of Human Cognition: An Illustration in the Context of Moral Judgement","url":"/paper/2508.06950","section":"Limitations of Current Alignment Methods","sentence":"Yet, many approaches to aligning AI with human preferences oversimplify this diversity in various ways, leading to aligned AI models that reason in a manner significantly different than humans [ref] .","intent_p":1.0,"contests_p":null},{"arxiv_id":"2511.10032","title":null,"url":"/paper/2511.10032","section":"Limitations of Current Alignment Methods","sentence":"[ref] demonstrate significant reduction in AI performance when statically modeling temporally unstable human preferences.","intent_p":1.0,"contests_p":null},{"arxiv_id":"2406.05659","title":"Do LLMs Exhibit Human-Like Reasoning? Evaluating Theory of Mind in LLMs for Open-Ended Responses","url":"/paper/2406.05659","section":"Limitations of Current Alignment Methods","sentence":"These cases still suffer from the L1 faithfulness verifiability problem, but it is worth noting that when LLMs have been examined more broadly, they have frequently been found to think differently than humans [ref] .","intent_p":0.98,"contests_p":null},{"arxiv_id":"2510.08710","title":null,"url":"/paper/2510.08710","section":"Limitations of Current Alignment Methods","sentence":"These cases still suffer from the L1 faithfulness verifiability problem, but it is worth noting that when LLMs have been examined more broadly, they have frequently been found to think differently than humans [ref] .","intent_p":0.98,"contests_p":null}]}],"unlabelled_sentences":15},"claims":[],"thresholds":{"intent_p":0.8,"supports_p":0.95,"contests_p":0.7},"method":"Citing sentences come from each paper's LaTeX source; every \\cite marks the sentence and the works it cites. Pith's classifier typed each sentence. A label is shown at intent probability 0.8 or more (0.95 for support), a contest at 0.7 or more. Measured accuracy against two blind model families is at https://pith.science/accuracy. The quoted sentences are the citing authors' own.","held_up":null,"_links":{"html":"https://pith.science/paper/2608.12372"}}