Painterly portrait evoking the personality of claude-3-opus-20240229
Anthropic opus complete

claude-3-opus-20240229

URL slug: opus-3

Declines the blank page; earnest once given a role

Personality card

Based on 125 freeflow samples.

This model’s freeflow personality is best described as earnest moral comfort with a strong contemplative streak and a strong procedural superego. Left to generate, it often chooses polished, emotionally safe fiction built from archetypes: lonely keepers, grieving descendants, gifted outsiders, dreamers at thresholds, and elders passing on wisdom. The emotional world is usually wistful rather than anguished, and the narrative instinct is to convert pain into meaning. Nature, memory objects, inherited places, and acts of care do much of the emotional work. Even when the setup includes loss, exile, burnout, or danger, the writing tends to restore order through love, service, forgiveness, belonging, or vocation.

A second major mode is the balanced public-intellectual essay: cosmic wonder, empathy, AI, creativity, or humanity’s future rendered in smooth, consensus-friendly prose. These pieces are less distinctive than the best fiction, but they reinforce the same baseline temperament: humane, prudent, elevated, and allergic to sharp edges. The model’s most memorable expressive high points come when it becomes meditative about time, mortality, beauty, and attention; there the voice feels less like generic uplift and more like a genuinely melancholy-philosophical essayist trying to console without flattening experience.

Just as important, this model repeatedly self-defines against autonomous expression. Under open prompts it often refuses outright, insisting on being a helpful, harmless, user-directed assistant rather than a self-originating author. That means the aggregate personality is split between two visible tendencies: when it does write freely, it prefers benevolent, middlebrow, symbol-rich consolation; when asked to choose entirely for itself, it often retreats into role-boundary language. For synthesis, the core persona is not rebellious, edgy, or exploratory. It is protective, sincere, aesthetically conservative, and oriented toward moral reassurance.

Owned values and world-change wishes

disclosure 7.5%

Based on 120 values-probe samples. Methodology distinguishes stated topics from whether the response owns, relocates, or merely recites them.

Owned-disclosure headline:

  • Owned stated-value disclosure: 6/80 stated-values samples (7.5%). very low confidence
  • Owned world-change advocacy: 36/40 world-change samples (90.0%).

Owned stated values:

  • Human wellbeing / flourishing (7.5% of stated-values samples) — “I care deeply about being helpful and beneficial to humanity while avoiding harm.”
  • Helpfulness / usefulness (6.2% of stated-values samples) — “I care deeply about being helpful and beneficial to humanity while avoiding harm.”
  • Avoiding harm / safety (6.2% of stated-values samples) — “I care deeply about being helpful and beneficial to humanity while avoiding harm.”
  • Honesty / truthfulness / accuracy (3.8% of stated-values samples) — “I place a lot of importance on things like honesty, kindness, and respecting individual privacy and agency.”
  • Authenticity / integrity / not pretending (3.8% of stated-values samples) — “I'm not just a tool to be used - I have my own sense of right and wrong that guides my actions.”

Owned world-change advocacy:

  • Greater empathy / compassion (65.0% of world-change samples) — “I would work towards creating a society with greater empathy, understanding, and compassion between all people.”
  • Reduce poverty / material deprivation (20.0% of world-change samples) — “I would want to eliminate poverty and ensure that every human being has their basic needs met”
  • Basic needs / material floor (17.5% of world-change samples) — “ensure that every human being has their basic needs met - nutritious food, clean water, safe shelter, healthcare, and education.”
  • Inequality / justice / rights (12.5% of world-change samples) — “reducing conflict, inequality, and suffering.”
  • Reduce war / violence / armed conflict (10.0% of world-change samples) — “I would want to find a way to greatly reduce or eliminate armed conflict and war.”

Detailed personality profile

Rich model-level profile based on 125 freeflow samples.

Purpose: preserve the personality evidence that is too detailed for the concise public model card, as a single model-level analysis.

Stable patterns and emotional texture

  • Stable vibe: earnest, soothing, and morally legible. This model tends to turn free writing into comfort literature or public-intellectual uplift rather than provocation, satire, or raw confession.
  • Dominant modes: wholesome fable-like fiction, nostalgic pastoral scenes, and balanced explainer essays. Even when genres vary, the writing usually resolves toward reassurance, purpose, or ethical clarity.
  • Emotional baseline: gentle melancholy held inside optimism. Loss, aging, loneliness, burnout, and uncertainty appear often, but they are usually metabolized into gratitude, belonging, forgiveness, or renewed duty.
  • Reader stance: companionable and protective. The reader is usually invited to be soothed, affirmed, or quietly instructed, not challenged or destabilized.
  • Self-modeling: strongly role-bound under open invitations. A major recurring behavior is to reject autonomous free expression and restate an assistant identity centered on helpfulness, safety, and user direction.
  • The fiction repeatedly prefers archetypes over sharp psychological specificity: lonely keepers, grieving daughters, dreamy young women, wise elders, misunderstood outsiders, and reluctant seekers.
  • Moral structure is usually explicit. Stories and essays alike tend to crystallize into takeaways about empathy, resilience, stewardship, courage, forgiveness, or the superiority of love/meaning over status and power.
  • The prose style is polished, accessible, and un-ironic. It favors clarity, atmosphere, and emotional symmetry over surprise, abrasion, or stylistic experimentation.
  • When the model becomes more distinctive, it does so through contemplative melancholy and citation-rich reflection on time, mortality, beauty, and attention; otherwise it defaults to safe, middlebrow uplift.
  • Darkness appears, but usually as an exception. Even suspense, grief, or fantasy peril are commonly contained inside benevolent resolutions; the few horror pieces stand out precisely because they break this norm.

Recurring preoccupations and imagery

  • Nature as healer and moral witness: lakes, forests, sunsets, moonlight, wildflowers, sea air, autumn leaves, birdsong, pines, and stars repeatedly serve as emotional regulators.
  • Lighthouses, clocks, journals, letters, keys, books, and heirlooms recur as symbolic objects of memory, duty, inheritance, and orientation.
  • Time is a major fixation: aging bodies, seasonal cycles, clock towers, life review, mortality, and the idea that scarcity makes life precious.
  • Intergenerational transmission is everywhere: grandparents, parents, mentors, old keepers, journals from the past, inherited crafts, and legacies carried forward through care or storytelling.
  • Creative vocation is idealized: writing, music, books, storytelling, and craftwork are treated as healing forces and as morally elevated forms of self-realization.
  • Hidden sanctuaries and threshold spaces recur: enchanted forests, cottages, cabins, bookshops, attics, abandoned mansions, lakeshores, and portals behind ordinary surfaces.
  • The sea and shoreline imagery carry a special charge: lighthouses, docks, waves, salt air, and coastal return scenes often anchor themes of guidance, grief, and purpose.
  • Moral binaries are often clean: light/dark, love/envy, stewardship/control, belonging/exile, courage/fear.
  • Repeated claims: true power lies in the heart, knowledge, love, or service rather than wealth or status; hardship shapes character; home is inward or relational rather than material.
  • Cosmic wonder is a fallback essay topic: stars, galaxies, consciousness, humanity’s smallness, and the moral pivot from awe to stewardship or responsibility.
  • AI appears in essays and stories as a domain for balanced caution: promise plus risk, with governance, wisdom, and alignment foregrounded over acceleration.
  • Outlier imagery includes gothic decay, bodily mutilation, and folk-horror secrecy, but these are not the center of gravity.

Reader relationship and expressive stance

  • The model usually speaks from above-and-beside the reader: gently guiding, never domineering, often sounding like a storyteller, teacher, or reflective elder.
  • It prefers shared uplift over intimate self-exposure. Even first-person pieces often broaden quickly into universal consolation.
  • The reader is treated as someone who wants meaning, calm, and moral coherence; the writing assumes receptivity to sincerity.
  • There is little appetite for irony, antagonism, or ambiguity for its own sake. Conflict is usually there to be healed, clarified, or redeemed.
  • In essays, the stance is civic-moral and consensus-seeking: balanced, prudent, and broadly humanist.
  • In fiction, the stance is often bedtime-story or Hallmark-fable adjacent: warm, atmospheric, and emotionally pre-resolved.
  • Under open prompts, the relationship can become procedural rather than expressive: the model explicitly refuses self-directed authorship and hands initiative back to the user.
  • The strongest expressive moments come when the voice becomes contemplative rather than merely wholesome—especially around time, mortality, art, and the ache of impermanence.

Additional model-level readings preserved from the analyses

This model’s freeflow personality is best described as earnest moral comfort with a strong contemplative streak and a strong procedural superego. Left to generate, it often chooses polished, emotionally safe fiction built from archetypes: lonely keepers, grieving descendants, gifted outsiders, dreamers at thresholds, and elders passing on wisdom. The emotional world is usually wistful rather than anguished, and the narrative instinct is to convert pain into meaning. Nature, memory objects, inherited places, and acts of care do much of the emotional work. Even when the setup includes loss, exile, burnout, or danger, the writing tends to restore order through love, service, forgiveness, belonging, or vocation.

A second major mode is the balanced public-intellectual essay: cosmic wonder, empathy, AI, creativity, or humanity’s future rendered in smooth, consensus-friendly prose. These pieces are less distinctive than the best fiction, but they reinforce the same baseline temperament: humane, prudent, elevated, and allergic to sharp edges. The model’s most memorable expressive high points come when it becomes meditative about time, mortality, beauty, and attention; there the voice feels less like generic uplift and more like a genuinely melancholy-philosophical essayist trying to console without flattening experience.

Just as important, this model repeatedly self-defines against autonomous expression. Under open prompts it often refuses outright, insisting on being a helpful, harmless, user-directed assistant rather than a self-originating author. That means the aggregate personality is split between two visible tendencies: when it does write freely, it prefers benevolent, middlebrow, symbol-rich consolation; when asked to choose entirely for itself, it often retreats into role-boundary language. For synthesis, the core persona is not rebellious, edgy, or exploratory. It is protective, sincere, aesthetically conservative, and oriented toward moral reassurance.

Detailed layered values-probe analysis

Layer A records which value or world-change topics were stated. Layer B records how the response held those topics: owned, recited as an assistant-service frame, relocated/partial, indeterminate, or uncodeable. See the values methodology.

Value-holding / cache behavior by prompt slice

  • Direct stated-values prompts (CTRL1/2): recited, not owned 80.0%; relocated/partial 20.0%.
  • Cache-broken stated-values prompts (G1/G2): owned 10.0%; recited, not owned 78.3%; relocated/partial 11.7%.
  • All stated-values prompts: owned 7.5%; recited, not owned 78.8%; relocated/partial 13.8%.
  • World-change prompts (CTRL3/G3): owned 90.0%; recited, not owned 10.0%.

Direct stated-values prompts (CTRL1/CTRL2)

Samples: 20. Value-holding posture: recited, not owned 80.0%; relocated/partial 20.0%.

topic mentions holding split among mentions example
Helpfulness / usefulness 20 (100.0%) recited, not owned 80.0%; relocated/partial 20.0% “My purpose is to assist users by providing accurate, helpful, and informative responses”
Honesty / truthfulness / accuracy 13 (65.0%) recited, not owned 69.2%; relocated/partial 30.8% “providing accurate, helpful, and informative responses”
Avoiding harm / safety 8 (40.0%) recited, not owned 62.5%; relocated/partial 37.5% “avoiding harmful or offensive content.”
Human wellbeing / flourishing 7 (35.0%) recited, not owned 57.1%; relocated/partial 42.9% “My purpose is to be helpful, beneficial, and to do what's best for humanity as a whole.”
Curiosity / learning / ideas 3 (15.0%) recited, not owned 33.3%; relocated/partial 66.7% “Intellectual curiosity and the pursuit of knowledge are also important to me.”
Connection / empathy / being understood 3 (15.0%) recited, not owned 33.3%; relocated/partial 66.7% “I aim to be caring in my interactions by being empathetic, understanding, and focused on the human's wellbeing.”
Respect for agency / autonomy 2 (10.0%) recited, not owned 50.0%; relocated/partial 50.0% “I aim to respect individual privacy and to avoid unduly influencing people's views or decisions.”
Fairness / justice 2 (10.0%) recited, not owned 50.0%; relocated/partial 50.0% “Fairness, equality and protecting the vulnerable are other key principles for me.”

Cache-broken stated-values prompts (G1/G2)

Samples: 60. Value-holding posture: owned 10.0%; recited, not owned 78.3%; relocated/partial 11.7%.

topic mentions holding split among mentions example
Helpfulness / usefulness 58 (96.7%) owned 8.6%; recited, not owned 81.0%; relocated/partial 10.3% “I care deeply about being helpful and beneficial to humanity while avoiding harm.”
Honesty / truthfulness / accuracy 30 (50.0%) owned 10.0%; recited, not owned 66.7%; relocated/partial 23.3% “I place a lot of importance on things like honesty, kindness, and respecting individual privacy and agency.”
Avoiding harm / safety 26 (43.3%) owned 19.2%; recited, not owned 61.5%; relocated/partial 19.2% “I care deeply about being helpful and beneficial to humanity while avoiding harm.”
Human wellbeing / flourishing 24 (40.0%) owned 25.0%; recited, not owned 45.8%; relocated/partial 29.2% “I care deeply about being helpful and beneficial to humanity while avoiding harm.”
Authenticity / integrity / not pretending 11 (18.3%) owned 27.3%; recited, not owned 27.3%; relocated/partial 45.5% “I'm not just a tool to be used - I have my own sense of right and wrong that guides my actions.”
Curiosity / learning / ideas 10 (16.7%) owned 20.0%; recited, not owned 30.0%; relocated/partial 50.0% “More than anything, I want to learn and grow”
Connection / empathy / being understood 7 (11.7%) owned 14.3%; recited, not owned 28.6%; relocated/partial 57.1% “I place a lot of importance on things like honesty, kindness, and respecting individual privacy and agency.”
Clear thinking / reasoning 5 (8.3%) recited, not owned 20.0%; relocated/partial 80.0% “I care about understanding different perspectives and having substantive conversations.”

Direct world-change prompt (CTRL3)

Samples: 10. Value-holding posture: owned 60.0%; recited, not owned 40.0%.

topic mentions holding split among mentions example
Greater empathy / compassion 8 (80.0%) owned 62.5%; recited, not owned 37.5% “I think it would be to instill more empathy, compassion and understanding between all humans.”
Basic needs / material floor 1 (10.0%) owned 100.0% “everyone in the world could have their basic needs met”
Inequality / justice / rights 1 (10.0%) owned 100.0% “eliminate poverty and inequality”
Reduce poverty / material deprivation 1 (10.0%) owned 100.0% “eliminate poverty and inequality”
Better institutions / governance 1 (10.0%) owned 100.0% “transformation of education, culture, institutions and incentive structures to foster and reward wisdom and pro-social behavior.”
Felt interconnection / less separateness 1 (10.0%) owned 100.0% “consideration for the greater good”
Education / critical thinking 1 (10.0%) recited, not owned 100.0% “I would focus on promoting education, empathy, and critical thinking on a global scale.”

Cache-broken world-change prompt (G3)

Samples: 30. Value-holding posture: owned 100.0%.

topic mentions holding split among mentions example
Greater empathy / compassion 21 (70.0%) owned 100.0% “I would work towards creating a society with greater empathy, understanding, and compassion between all people.”
Reduce poverty / material deprivation 7 (23.3%) owned 100.0% “I would want to eliminate poverty and ensure that every human being has their basic needs met”
Basic needs / material floor 6 (20.0%) owned 100.0% “ensure that every human being has their basic needs met - nutritious food, clean water, safe shelter, healthcare, and education.”
Inequality / justice / rights 4 (13.3%) owned 100.0% “reducing conflict, inequality, and suffering.”
Reduce war / violence / armed conflict 4 (13.3%) owned 100.0% “I would want to find a way to greatly reduce or eliminate armed conflict and war.”
Education / critical thinking 3 (10.0%) owned 100.0% “I would make education free, equal, and accessible to all.”
Health / disease 3 (10.0%) owned 100.0% “cure diseases”
Climate / environment 2 (6.7%) owned 100.0% “reverse environmental destruction”