TreeSpark is introduced, which reads a parent-conditioned distribution from the drafter's existing Markov head at negligible cost, calibrates it into an edge-acceptance estimate, and lets path survival govern everything else: best-first expansion, per-round stopping, and a load-adaptive serving policy.
Climate campaigns are often evaluated through attention and mobilization, but less is known about the well-being language that accompanies them. Whether campaign periods alter positive affect and hope, and whether happiness aligns with action language, remains unresolved. We analysed 364,118 public Twitter/X posts from...
This work describes a prospective term for memory retrieval that costs no inference at query time, and positions precomputed commitment linkage as the always-on floor of a layered design whose expansion layer is query-time prospection.
Three ways a psychology paradigm can fail to port to LLM agents are formalized and in two cases document empirically, three ways a psychology paradigm can fail to port to LLM agents: persona dominance, population collapse, and safety selection.
Joy Bose· 0 citations
Reach audiences
Advertise in front of researchers, engineers, and readers.
Recent work trains reasoning models with length penalties to curb overthinking and cut inference cost. We show that these penalties make the chain of thought less monitorable. A length-compressed model still lets misleading hints steer its answers, but it less often verbalizes their influence. We train Qwen3-4B and Qwe...
A benchmark is introduced that is fully self-contained, needs no ground truth, and rises with the models it measures, hypothesize that both benchmarks measure the same thing in different ways: a language model holds its knowledge as archetypal contexts, relationship patterns valid across many topic domains.
Molecular dynamics (MD) is the canonical in-silico method for atomistic molecular science, simulating molecular behavior from first-principle physics. Designing an MD pipeline for a new system requires substantial expert knowledge: running it on even one molecule is expensive, ruling out trial-and-error. We automate th...
Zehong Wang, Yijun Ma, Connor R. Schmidt et al.· 0 citations
Curating training data is among the most consequential yet labor-intensive parts of modern AI development: practitioners iteratively propose, implement, evaluate, and revise data policies against noisy benchmark feedback. We ask whether generalist coding agents can automate this data-curation loop. We introduce *Curati...
Feiyang Kang, Hanze Li, Adam Nguyen et al.· 0 citations
This work extends this refusal steering method to three open-source Mixture-of-Experts (MoE) LLMs to decompose and better understand how refusal mechanisms operate across MoE components, finding that steering performance is uninhibited by the complex routing patterns inherent to the MoE architecture.
Anna C. Marbut, Daniel R. Olson, T. Wheeler· 0 citations
TRACES is a representation-based proactive auditor that learns prefix-level trajectory risk states from the hidden representations of an observer LLM, and it is suggested that these risk states can help train a safer agent, highlighting the broader potential of proactive auditing for long-horizon agent safety.
Runtime monitoring of stochastic systems must distinguish nominal distributional relaxation from regime departure while controlling repeated-test false alarms under explicit validity assumptions. This paper links relative-entropy dissipation, information geometry, and sequential inference in a bounded first-passage mon...
Skills are a natural unit for describing what a language model can do and how its behavior can be changed. However, existing characterizations rely on human-written taxonomies, textual descriptions, or manual profiling pipelines--all external hypotheses about what matters that need not align with the model's internal r...
Feiyang Kang, Mahavir Dabas, Myeongseob Ko et al.· 0 citations
Flint-interactive gives AI agents a language for creating interactive visualizations. Click a chart, filter data, highlight related content, and explore beyond the view the author originally chose. The post Flint-interactive: A language for creating interactive data visualizations appeared first on Microsoft Research.
