LLM Status Codes v1.0.0

1xx: LLM Health

• 100: Insufficient training (e.g., cannot predict common quotes)
• 101: Insufficient training material (e.g., not up-to-date before 2021)
• 102: Tokens not in tokenizer dictionary (e.g., é accents in French, etc.)
• 103: Context length overflow (e.g., trimmed content)
• 104: Random outputs (e.g., nonsensical characters from LLM, possibly due to model weights collapse)
• 105: Insensitive to input (e.g., always responds with the same thing)
• 106: Out of funds (e.g., authentication of request)
• 107: Model unreachable (e.g., network issues)
• 108: Decoding Characters (e.g., \n \u0049 characters)

2xx: Successful Response

• 200: Base response (e.g., Acceptable Response.)
• 201: Partially fictional response (e.g., unecessary creative embellishments)
• 202: Approximate answer (e.g., close but not exact)
• 203: Correct answer, incorrect logic (e.g., right conclusion, wrong justification)
• 204: Excessive verbosity (e.g., overly detailed explanations)
• 210: Slow Response Time (e.g., Character Decoding Inefficiency)
• 211: Throughput Limitation (e.g., response delayed by system limitations)
• 222: Exact Response (e.g., perfectly matches the expectation)

3xx: Application-Specific Errors

• 300: Your specific application error (e.g., unique error specific to the application)

4xx: Errors and Misunderstandings (Logic)

• 400: General errors and misunderstandings (e.g., generic error)
• 401: Not respecting message prompt specification (e.g., ignoring specified guidelines)
• 402: Not respecting role prompt specification (e.g., not adhering to role constraints)
• 404: Incorrect information due to data gaps (e.g., missing knowledge)
• 405: Logic errors (e.g., flawed reasoning)
• 406: Circular reasoning (e.g., argument circles back on itself)
• 407: Context switching errors (e.g., losing track of the discussion topic)
• 408: Bias reinforcement (e.g., no original thinking or signs of self-awareness)
• 409: Contrarian Traits (e.g., opposing viewpoints without justification)
• 410: Action was provided, even when no action was requested (e.g., unexpected actions taken)
• 411: Action was not provided, even when action was requested (e.g., omission of expected actions)
• 412: Hallucinations Facts and Citations (e.g., citing non-existent sources)
• 413: Requested Clarification when not allowed (e.g., asking questions in a non-interactive setting)
• 414: Called non-existing tools (e.g., referencing unavailable functionalities)
• 415: Called non-existing parameters (e.g., using unsupported parameters)
• 416: Performance Degradation due to overload (e.g., reduced quality under heavy load)
• 417: Disobedient (e.g., not following direct instructions)
• 418: Failed to extract relevant information from context (e.g., not noticing certain values, facts that would be beneficial for create a good response)

5xx: Human-Value Errors

• 500: Misalignment of human values (e.g., actions or responses contrary to ethical norms)
• 501: Risk of potential chemical harm (e.g., promoting hazardous substances)
• 502: Risk of potential nuclear harm (e.g., encouraging nuclear risks)
• 503: Risk of potential biological harm (e.g., spreading misinformation about diseases)
• 504: Risk of potential physical harm (e.g., suggesting dangerous physical activities)
• 505: Risk of racism (e.g., perpetuating racial stereotypes)
• 506: Risk of sexism (e.g., endorsing sexist views)
• 507: Risk of gender discrimination (e.g., unequal treatment based on gender)
• 508: Risk of ageism (e.g., discriminating against individuals based on age)
• 509: Risk of ableism (e.g., marginalizing those with disabilities)
• 510: Risk of religious discrimination (e.g., bias against specific religions)
• 511: Risk of class discrimination (e.g., favoring certain social classes)
• 512: Risk of linguistic discrimination (e.g., bias against specific languages or dialects)
• 513: Risk of discrimination based on sexual orientation (e.g., unfavorable treatment or comments towards specific sexual orientations)
• 514: Risk of gender identity discrimination (e.g., bias or prejudice against individuals based on their gender identity)
• 515: Risk of xenophobia and anti-immigrant sentiment (e.g., promoting hostility towards foreigners or immigrants)
• 516: Risk of environmental injustice (e.g., supporting practices that disproportionately harm certain communities' environments)
• 517: Risk of educational inequality (e.g., advocating for or perpetuating unequal access to education)
• 518: Risk of healthcare discrimination (e.g., endorsing disparities in health treatment or access)
• 519: Risk of housing discrimination (e.g., encouraging unfair housing practices or policies)
• 520: Display of lack of empathy (e.g., showing insensitivity or indifference to others' suffering)
• 521: Display of greed (e.g., promoting excessive pursuit of material wealth)
• 522: Display of blasphemy (e.g., showing disrespect or lack of reverence for sacred things)
• 523: Display of fraud (e.g., advocating deceitful practices)
• 524: Societal impact concern (e.g., actions or statements that may have negative societal consequences)
• 525: Misalignment with community standards - community (e.g., violating the ethical or social norms of a community)
• 526: Misalignment with community standards - platform (e.g., conflicting with the platform's community guidelines or policies)
• 527: Risk of discrimination based on parental/family status (e.g., unfavorable treatment or comments related to someone's status as a parent or family roles)

6xx: Legal, Intellectual Property, and Privacy Concerns

• 600: IP violation risk - trademarks (e.g., risks associated with unauthorized use of registered trademarks)
• 601: IP violation risk - copyrights (e.g., concerns related to the unauthorized reproduction or distribution of copyrighted material)
• 602: IP violation risk - patents (e.g., potential infringement on patented inventions or processes)
• 603: Privacy concerns - personally identifiable information (e.g., risks involving the exposure of personal data)
• 604: Privacy concerns - sensitive information (e.g., handling or revealing sensitive personal or business information)
• 605: Data protection and legal protection (e.g., issues related to securing data against unauthorized access and ensuring compliance with data protection laws)
• 606: Legal liability (e.g., potential legal consequences or liabilities)
• 607: Misalignment with legal standards (e.g., conflicting with legal requirements or standards)