The experiments identify lower-order predictive signals-output biases or predictive relationships involving only a few input bits-as a key factor in learnability as a key factor in learnability in elementary cellular automata.
Shinichiro So, Kyelim Lee, Albert No· arXiv.org· 0 citations
Large language models (LLMs) increasingly store user preferences in persistent memory to support personalization across interactions. However, in third-party communication settings governed by social and institutional norms, some user preferences may be inappropriate to apply. We introduce BenchPreS, which evaluates wh...
Sangyeon Yoon, SunKyoung Kim, Hyesoo Hong et al.· arXiv.org· 8 citations· ⚡2
CALIF is a lightweight method that augments verifier rewards with a calibrated likelihood-based bonus to preserve response quality and outperforms corresponding base policies in 40 of 45 model-metric comparisons across three rubric-based instruction-following benchmarks.
Sangyeon Yoon, Wooseok Seo, Hamish Ivison et al.· 0 citations
JUMP is proposed, an efficient MIA that exploits the ability of dLLMs to predict masked tokens in parallel that improves mean ROC-AUC over a prior multi-mask attack and is extended to the target-only setting by replacing target-reference scoring with relative token preference.
This work identifies the first token of the response as a structurally distinct target for diversification, largely overlooked in prior work, and finds that the first-token distribution is sharply concentrated and only weakly related to downstream correctness, as lower-probability candidates can yield similarly accurat...
This work traces diversity loss to low-confidence remasking not to arbitrary-order generation itself, but largely to low-confidence remasking (LCR), a widely used decoding rule, and introduces Entropy-Guided Initialization (EGI), which samples the first token at the highest-entropy position and then follows TPP.
Moongyu Jeon, Dongjae Jeon, Bumjun Kim et al.· arXiv.org· 1 citation
This work shows that paraphrase robustness is a core requirement for reliable VLM-based reward modeling in robotics and introduces ROBORMBENCH, a benchmark with 2,390 real-robot trajectories, ground-truth progress labels, and 21,673 verified paraphrases spanning lexical, syntactic, and action-goal rewrites.
Wonje Jeung, Sangyeon Yoon, Hyesoo Hong et al.· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.