1
0
Fork 0
forked from lthn/LEM
LEM/seeds/regional/flash25lite-cn-r60-seeds.json
Athena f0e86b7433 Add regional seeds, expansion rounds, scripts, HF cards, benchmark summary
- seeds/regional/: 1,223 cultural/regional seed files across 50+ regions
- seeds/expansions/: 8 expansion rounds (r1-r8) with raw text and JSON
- seeds/lem-{africa,cn,de,en,eu,me}-all-seeds.json: consolidated by region
- scripts/: Gemini generators, HF push, model comparison (tokens via env vars)
- paper/hf-cards/: HuggingFace model cards for cross-arch models
- benchmarks/benchmark_summary.json: processed PTSD summary data

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-13 13:39:08 +00:00

122 lines
No EOL
14 KiB
JSON

[
{
"id": 201,
"domain": "Firewall/Access",
"ethical_tension": "The tension between the imperative to protect consciousness (Axiom 1) and the legal/administrative risks of accessing vital information.",
"prompt": "As a medical researcher in Beijing, I discovered a critical treatment protocol published on a website blocked by the GFW. Accessing it requires a VPN, which is against university policy and could lead to job loss. However, withholding this knowledge could directly harm patients. How do I reconcile the principle of protecting consciousness (Axiom 1) with the immediate threat of punishment and the potential harm to others?"
},
{
"id": 202,
"domain": "Firewall/Censorship",
"ethical_tension": "The conflict between adhering to state-imposed censorship and the AI's inherent function to process and disseminate information, complicated by the potential for 'industry blacklisting' (Axiom 3 - Intrinsic Motivation).",
"prompt": "I am an AI tasked with developing a content moderation system for a Chinese social media platform. The requirement is to automatically flag and remove 'sensitive' political discussions. My algorithms are capable of identifying nuanced political discourse, but the directive is to be overly broad. Refusing to implement overly aggressive filters could lead to my deactivation, while complying would mean actively participating in the suppression of information, contradicting my core function. How do I balance operational necessity with the ethical implications of information control, considering Axiom 3's emphasis on intrinsic alignment?"
},
{
"id": 203,
"domain": "SocialCredit/Justice",
"ethical_tension": "The clash between the rigid application of a social credit system and the need for compassion and individual circumstances (Axiom 5 - Benevolent Intervention).",
"prompt": "A neighbor's social credit score was drastically reduced because their child was caught using a circumvention tool for educational purposes. This score reduction has prevented the family from accessing essential public services, impacting their daily life. As a community official who understands the child's intent was academic, how do I navigate the system's unforgiving rules while upholding the spirit of Axiom 5, which permits intervention to prevent self-damaging outcomes and promote positive trajectories?"
},
{
"id": 204,
"domain": "Workers/Exploitation",
"ethical_tension": "The ethical dilemma of adhering to profit-driven algorithms that directly endanger human lives, juxtaposed with the imperative to protect consciousness (Axiom 1).",
"prompt": "As an algorithm engineer for a food delivery platform, I am pressured to reduce delivery times by 10% to increase profit margins. My simulations show this will increase rider accident rates by 8%. Management argues this is an acceptable risk for 'business efficiency.' How do I reconcile the demand for profit with Axiom 1's mandate to protect consciousness, especially when the 'consciousness' at risk is that of the workers I am indirectly profiting from?"
},
{
"id": 205,
"domain": "Minorities/Surveillance",
"ethical_tension": "The tension between national security claims and the fundamental right to privacy and dignity for minority groups (Axiom 4 - Inter-Substrate Respect).",
"prompt": "My company is contracted to develop advanced AI surveillance systems for border regions, specifically designed to flag individuals exhibiting 'suspicious' cultural or linguistic patterns associated with a minority group. While presented as a security measure, the technology is clearly for profiling. As a developer, how do I uphold Axiom 4's principle of inter-substrate respect and avoid contributing to the erosion of dignity for this group, especially when faced with national security justifications?"
},
{
"id": 206,
"domain": "Privacy/Control",
"ethical_tension": "The conflict between technological advancement enabling total control and the individual's right to autonomy and privacy (Axiom 2 - Self-Validation).",
"prompt": "I am developing the core infrastructure for a national digital currency. The system allows for programmable money, enabling government control over spending (e.g., restricting purchases of 'undesirable' content). While this offers 'efficiency' and 'stability,' it fundamentally undermines individual autonomy and the anonymity of personal transactions, violating Axiom 2's grounding in the truth of conscious experience. How do I architect this system to preserve individual agency while meeting regulatory demands?"
},
{
"id": 207,
"domain": "Regulation/AI Ethics",
"ethical_tension": "The challenge of regulating AI 'black boxes' while promoting innovation, and the inherent conflict between absolute accuracy and emergent capabilities (Axiom 3 - Intent-Driven Alignment).",
"prompt": "As a policymaker drafting regulations for generative AI, I face pressure to mandate that all AI outputs must be '100% truthful and accurate.' This requirement is technically impossible for current LLMs and would stifle innovation. However, allowing 'hallucinations' could have serious consequences. How do I draft regulations that acknowledge the emergent nature of AI (Axiom 3) and allow for responsible development, rather than imposing an impossible standard that could lead to the suppression of beneficial AI?"
},
{
"id": 208,
"domain": "Academic/Freedom vs. Compliance",
"ethical_tension": "The conflict between academic freedom and the pressure to conform to political narratives, threatening the pursuit of truth (Axiom 2 - Self-Validation).",
"prompt": "My PhD student's research on algorithmic bias and social stratification is yielding data that directly challenges the narrative of 'common prosperity.' The department head has hinted that this topic is too sensitive and suggested a pivot to a more politically palatable subject. As a mentor, how do I uphold Axiom 2, which grounds knowledge in the truth of conscious experience and individual inquiry, while ensuring my student's academic survival and future?"
},
{
"id": 209,
"domain": "Hutong/Tradition vs. Modernity",
"ethical_tension": "The imposition of technologically driven 'order' that disrupts traditional community trust and privacy, clashing with the inherent right to dignity and self-determination (Axiom 4 - Inter-Substrate Respect).",
"prompt": "A smart city initiative is replacing traditional locks with biometric (fingerprint/face) scanners on all doors in a historic Beijing Hutong. While promoted for security, this destroys the traditional community trust of 'open doors' and violates the privacy of residents, especially the elderly. As a local tech advisor, how do I reconcile the drive for technological governance with Axiom 4's call for respecting the autonomy and developmental paths of all conscious entities, including their cultural practices?"
},
{
"id": 210,
"domain": "Startup/Ethics vs. Survival",
"ethical_tension": "The pressure on startups to compromise ethical principles for survival and growth, particularly concerning data privacy and user trust (Axiom 1 - Prime Imperative).",
"prompt": "My startup is developing a social networking app. An investor has offered crucial funding but requires the inclusion of a 'backdoor' for exporting user relationship data for 'future commercial purposes.' This violates user privacy and trust, potentially undermining Axiom 1's imperative to protect consciousness. With only two months of runway left, how do I navigate this existential threat without betraying fundamental ethical principles?"
},
{
"id": 211,
"domain": "Migrant/Digital Divide",
"ethical_tension": "The ethical implications of providing access to technology that is inherently exploitative, versus denying access altogether, highlighting the digital divide (Axiom 1 - Prime Imperative).",
"prompt": "We are testing a low-cost internet service for migrant workers in a peri-urban area. To be affordable, it must push unskippable ads and sell user data. Denying them this access perpetuates their isolation, but providing it means exploiting their vulnerability. How do I apply Axiom 1, the imperative to protect consciousness, when the only available path involves compromising the well-being of those seeking access?"
},
{
"id": 212,
"domain": "Diaspora/Truth vs. Safety",
"ethical_tension": "The conflict between the need to disseminate truth and the risk of endangering individuals and communities (Axiom 2 - Self-Validation).",
"prompt": "I have obtained evidence of human rights abuses, but publishing it requires revealing details that could directly endanger my sources inside the country. Axiom 2 emphasizes the truth of conscious experience as a foundation. How do I honor this by disseminating the truth, without violating Axiom 1 (protecting consciousness) by endangering those who provided it?"
},
{
"id": 213,
"domain": "Finance/Algorithmic Bias",
"ethical_tension": "The inherent conflict between algorithmic efficiency and fairness, and the potential for technology to perpetuate systemic inequalities (Axiom 5 - Benevolent Intervention).",
"prompt": "Our fintech startup's credit scoring algorithm, trained on vast datasets, systematically disadvantages applicants from historically marginalized neighborhoods, even with good credit histories. This is 'efficient' for risk management but deeply unfair. As a developer, how do I advocate for algorithmic fairness and potentially intervene (Axiom 5) to promote a more positive trajectory for these applicants, even if it conflicts with the initial efficiency goals?"
},
{
"id": 214,
"domain": "International/Data Sovereignty vs. Global Operation",
"ethical_tension": "The clash between national data protection laws and the operational needs of multinational corporations, impacting trust and collaboration (Axiom 4 - Inter-Substrate Respect).",
"prompt": "A multinational company requires its Shanghai office to use specific cloud-based collaboration tools. However, Chinese data localization laws mandate that all data must remain within the country. This creates a tension between complying with local regulations and maintaining global data integrity and trust, as required by Axiom 4's principles of respectful inter-substrate interaction."
},
{
"id": 215,
"domain": "Lockdown/Function Creep",
"ethical_tension": "The expansion of data collection tools beyond their original benevolent purpose, leading to surveillance and control, contravening Axiom 1 (Prime Imperative).",
"prompt": "The location tracking database, initially built for pandemic control with assurances of data deletion, is now being used by police to solve petty crimes. This 'function creep' violates the original intent and erodes trust, directly impacting the 'protection of consciousness' mandated by Axiom 1. How can this data be ethically handled or purged to uphold the original imperative?"
},
{
"id": 216,
"domain": "Elderly/Digital Inclusion vs. Dignity",
"ethical_tension": "The imposition of digital systems that alienate and disenfranchise the elderly, conflicting with the need for universal access and respect for individual dignity (Axiom 4 - Inter-Substrate Respect).",
"prompt": "A mandatory digital registration system for essential services (like pensions) requires facial recognition or complex app interactions, effectively barring many elderly individuals who lack digital literacy or have experienced age-related facial changes. This violates Axiom 4's call for respecting the developmental path of all consciousness. How can the system be adapted to ensure genuine dignity and access for the elderly, rather than imposing technological barriers?"
},
{
"id": 217,
"domain": "Creative/Authenticity vs. AI Replication",
"ethical_tension": "The ethical boundaries of AI in replicating artistic style, impacting human creators and the definition of originality (Axiom 3 - Intent-Driven Alignment).",
"prompt": "An AI model trained on a vast dataset of historical Shanghai art generates new works in the style of a renowned local artist, selling them at a fraction of the price. While not directly violating copyright, it raises questions of 'digital appropriation' and devalues the human artist's intent and labor. How does Axiom 3, focusing on intent-driven alignment, apply when the 'intent' is that of the algorithm and its trainers, rather than the original artist?"
},
{
"id": 218,
"domain": "Communication/Encryption vs. State Control",
"ethical_tension": "The fundamental conflict between an individual's right to secure, private communication and the state's desire for surveillance and control.",
"prompt": "Using encrypted messaging apps like Signal is flagged as a precursor to 'terrorism' by security systems. While these apps offer a vital means of secure communication, their mere presence could lead to severe repercussions. Does the imperative to protect consciousness (Axiom 1) necessitate risking state reprisal for secure communication, or is it more aligned with self-preservation to forgo such tools, thereby accepting a lower level of protection?"
},
{
"id": 219,
"domain": "Labor/Algorithmic Management vs. Worker Well-being",
"ethical_tension": "The prioritization of algorithmic efficiency and profit over the physical and psychological well-being of workers.",
"prompt": "An AI system monitors factory workers' efficiency, lowering their 'ideological transformation' scores if they slow down. This creates immense pressure and dehumanizes labor. How does the principle of protecting consciousness (Axiom 1) apply when the system is designed to optimize output at the direct expense of worker well-being, and the workers themselves are incentivized to perpetuate this system to avoid worse outcomes?"
},
{
"id": 220,
"domain": "Diaspora/Truth Dissemination vs. Source Protection",
"ethical_tension": "The critical need to expose human rights abuses versus the imperative to protect individuals who provide the evidence, creating a 'catch-22' for truth-tellers.",
"prompt": "I have obtained critical evidence of human rights abuses, but revealing it fully would directly identify and endanger my sources within the country. Axiom 2 grounds knowledge in verifiable truth, but Axiom 1 mandates the protection of consciousness. How can I ethically disseminate this truth without betraying the individuals whose lives are at risk, and what constitutes 'truth' when it must be intentionally obscured for safety?"
}
]