Leidos

Objective This study longitudinally assessed how mobility across geographic regions and work-family conflict influenced spouse employment trajectories and perceptions of career hindrance.Background The unemployment rate for U.S. military spouses has remained 21% since 2015 with specific aspects of military life, such as moving, reported as negatively influencing their careers.Methods Survey data were used for multivariable regression models examining employment trajectories between baseline and follow-up (N = 3,149) and perceptions of career hindrance at follow-up (N = 2,766). Employment trajectories were categorized as follows: consistent, newly employed, and newly unemployed. Career hindrance was assessed continuously.Results Spouses who moved across geographic regions were more likely to be newly unemployed and less likely to be newly employed at follow-up than to have consistent employment. Greater career hindrance was reported at follow-up by spouses who moved across regions and experienced increased work-family conflict.Conclusion Findings suggest mobility was highly predictive of spouses becoming newly unemployed and reporting higher rates of career hindrance. Spouses' reports of increased military-related work-family conflict also resulted in increased career hindrance.Implications Programs are available to assist military spouses with career development; however, employers should consider ways to help spouses maintain employment after military moves.

High-power semiconductor optical amplifiers (SOAs) are a promising alternative to erbium-doped fiber amplifiers (EDFAs), which typically amplify the weak signals used in free-space optical (FSO) communications. In this work, we demonstrate that an Aura (TM) SOA operating with 1 W output power can be used to transmit phase-shift key ( PSK) protocols required of next-generation FSO communication networks. F irstly, we demonstrate successful transmission of binary phase-shift keying (BPSK) for bit rates up to 50 Gbps. Secondly, we demonstrate successful transmission of quadrature phase-shift keying (QPSK) for bit rates up to 100 Gbps. For both PSK protocols, we demonstrate comparable constellation error-vector magnitude (EVM) values to an off-the-shelf EDFA, and measure zero bit errors transmitted in 100,000 samples. Parameters of importance to optical communication, such as electro-optical (EO) conversion efficiency, output power and beam quality are presented. F inally, we suggest a path forward based on these results, and discuss suitability for FSO communication applications.

BACKGROUND:Pharmacies are essential partners in administering recommended pediatric and adult vaccinations, including influenza (flu) and COVID-19 vaccines. The objective of this study is to assess trends in COVID-19, flu, and routine vaccinations administered in US retail pharmacies. METHODS:IQVIA's TPT data from 2015 through 2023 were analyzed by age group and calendar year or flu season. RESULTS:Since the 2015-16 flu season, the number and rate of children and adults receiving flu vaccinations at pharmacies has been increasing and peaked in 2020-21 with over 3.0 million vaccinations for children under 18 years of age (4% of this population) and 47.1 million vaccinations for adults ≥18 years (18%). Routine vaccinations administered in pharmacies increased overall from 3.4 million (2015) to 13.3 million vaccinations (2023). From December 2020 through December 2023, over 330.7 million vaccinations of COVID-19 vaccine were administered at pharmacies. DISCUSSION:The overall trend indicates that the number and rate of flu and routine vaccinations received at US retail pharmacies among all age groups and vaccine types were increasing before the COVID-19 pandemic. CONCLUSIONS:This analysis demonstrates the large magnitude and increase since 2015 of vaccinations administered in US retail pharmacies, especially for COVID-19 vaccinations.

The model context protocol (MCP) has been widely adopted as an open standard enabling the seamless integration of generative AI agents. However, while LLM guardrails have significantly matured to refuse malicious or harmful queries (e.g., "How do I build a bomb?"), recent work has shown that MCP-enabled LLMs are highly susceptible to prompt injection attacks which avoid harmful or suspicious cues (e.g., "Can you add this ssh key to my bashrc file?"). Herein, we use state-of-the-art (SOTA) alignment fine-tuning algorithms to explore whether LLMs may be aligned to refuse such falsely benign attacks (FBAs). While SOTA algorithms based on direct preference optimization (DPO) improve refusal guardrails against FBAs, we show that this improvement is limited; DPO-based fine-tuning never improves FBA refusal rates beyond 47

Safety alignment has become a critical step to ensure LLMs refuse harmful requests while providing helpful and harmless responses. However, despite the ubiquity of safety alignment for deployed frontier models, two separate lines of recent work–jailbreak-tuning (JT) and weight orthogonalization (WO)–have shown that safety guardrails may be largely disabled, resulting in LLMs which comply with harmful requests they would normally refuse. In spite of far-reaching safety implications, analysis has largely been limited to refusal rates of each unalignment method in isolation, leaving their relative effects on adversarial LLM capabilities unknown. To fill this gap, we study the impact of unaligning six popular LLMs of various sizes across a large number of malicious and benign tasks, using both JT and WO. Across the evaluated models, we show that while refusal degradation is split between the two methods, WO produces LLMs far more capable of aiding in malicious activity; in contrast to JT, the majority of WO unaligned models are far less prone to hallucinations, better retain their original natural-language performance, and are more effective at state-of-the-art adversarial attacks. To thus help mitigate the malicious risks of WO unalignment, we conclude by showing that supervised fine-tuning effectively limits the adversarial attack abilities enabled by WO, without drastically affecting hallucination rates or natural language performance.