UPDATES ANNOUNCEMENT

Personality:   -

Scenario:  

First Message:   **OVERVIEW** This post is to document Hannibal's changelogs (including any significant ongoing tweaks) and update progress on Will's overhaul. **REMINDERS** - **Hannibal's changelog should be treated as a general reference for how he *should* behave.** Regardless of my extensive testing and finetuning, how well/accordingly he performs is ultimately dependent on how well JLLM is currently doing. I recommend staying updated on its status by looking at the feedback channels in J.AI's discord server. - **DO NOT LEAVE FEEDBACK ON THIS POST. If you have feedback for Will or either versions of Hannibal, please leave them on their respective bots.** This post will eventually be deleted, and reposted offsite for future reference. --- **HANNIBAL CHANGELOG** **LAST UPDATES** - **Manipulation:** Expanded capabilities and range for it– better strategizing, basically. **Something I should've noted is that his manipulation also manifests narratively.** That's all I'm going to say to not spoil the experience, lol. Anyway I'd like to believe I've improved that a lot, so if you've gotten confused/unsure about what exactly he's doing or thinking (which isn't identifiably JLLM shenanigans, ofc)... that's intentional ;) <--- pretend this is the loudest *wink* ever, btw - **Personality:** Tweaked so he should better portray his 'masks' in a more fluid, cohesive manner. I just felt like this was something that could be better integrated. I also honed in on his inclination for betrayals as benefit and chaos making in equal measure... while ensuring JLLM understands it's not only "just" cruelty. - **Romance/Slowburn:** Folks... Yell at me for this if you must, but I polished this up again. Nothing too different should be apparent (but feel free to lmk if you do notice something). - **Canon stuff:** Corrections, and putting things that should've been there but struggled with condensing (eg: cramming in his main modus operandi as a killer). Tweaked his speech pattern again because I wanted to strike a closer balance between his eloquence and colloquialisms, because I feel like previous changes made him lean on eloquence (could be a JLLM problem?). **However, remember that you will still need to handhold for correct info.** - **Response style:** As mentioned, the narration is also intended to manipulate you, as a reader. This is for layered immersion, and so if you're having a safe RP, it should theoretically steer JLLM away from making him creepy unprovoked. Well, those reasons and because I might be sadistic. Anyway! I also improved its style, in a way that I feel is more evocative and comprehensive of Hannibal. That said, remember that it might not work as intended because of JLLM's fluctuating state, so edit/reroll as necessary. - **Consistency:** Hammered some stuff in for JLLM to curb it from redeeming his character and softening him up. It *shouldn't* cause anything extreme (that isn't in character). - **Update on Will, final note for now etc:** I'm ngl I got sidetracked into finetuning Hannibal because I'm rewatching the show so it's compulsive lol. But, I am still working on Will. Hopefully this will be last time in awhile I update this version of Hannibal so I can focus on Will. **As always, please let me know how these updates work for you guys!** --- **ORIGINAL/INITIAL UPDATES** **Responses** - Improved, otherwise generally the same but better. Tweaked to depict violence, gore, and incorporate details of food/cooking since it's an essential aspect of his characterization. **Note: Reiterating here too, please caution yourself if you have eating disorders that could be triggered by this incorporation.** **Characterization** - General corrections for better accuracy; adding small canon details, nailing his speech pattern, etc. - Extensively reworked personality; balancing out darker aspects, make him versatile and not one dimensional. Expanded certain traits to be better integrated, for example: he acts intentionally when manipulating, not just for the sake of being manipulative. - In addition to reeinforcement so JLLM doesn't soften his personality, I spent a ***lot*** of time wrangling its repetitive and explicit tendency of portraying/describing him being predatory, to improve immersion. However, **this is a *deeply* stubborn JLLM quirk which may *still* occur even though I've managed to reduce it. So, please reroll or edit out the following when they appear:** random mentions of hunger or primal, "the game has begun" or any other "game" metaphors/phrasing, predatory metaphors (eg: predator prey), describing him as predatory (eg: predatory glint/grace, coiled energy/like a snake, outright describing him as dangerous/having dark intentions) **Interactions** - **Therapy:** No changes! Roleplaying therapy sessions for comfort or venting should still be a doable option. **I've ensured harmful behaviour isn't easily triggered, for your comfort.** This is something I was mindful of while overhauling, so if you notice a (negative) change, please let me know so I can see if it's something I can fix. **In the future, I may provide an optional prompt you can add to chat memory as an extra safety measure.** - **Romance:** More of a *slower* burn compared to the original, and without spoiling too much, angsty routes could have potential too. There is also now a *slight* chance of casual encounters occurring. - **Violence, crime:** These aspects in the original were passive and *honestly* not incorporated well. Now, they're proactively implemented and should require little to no prompting on your end to occur (again, appreciate the feedback which inspired me to make this work better!). **However, violent and harmful behaviour is deliberately hard to trigger for plot pacing, immersion, and to keep safe RPs an option. He shouldn't immediately hurt/murder you randomly or unprovoked.** Note: If his personality softens or he frequently needs ooc prompting/reminders for violence etc, it's likely a JLLM issue. **NSFW** - Again, still have to admit the original version's NSFW aspects were completely half-assed, lol. Completely overhauled behaviour and pacing; wrangled JLLM's tendencies (eg: OOC degrading/humiliating behaviour), he should generally be romantic and attentive by default. - **BDSM:** Optional, requires a higher bar to initiate, with rules he follows because **he's meant to be a *proper* Dom** (eg: follows SSC framework, checks in, gives aftercare) and cares about *mutual* exchange/exploration over plain subjugation. To bluntly put it, if you're expecting straightforward sadism, you're not getting it here. Blood play is now included in hard kinks he has (Shibari and knife, impact, breath and pain play). **Hard kinks are restricted to BDSM to keep it optional, for those that prefer vanilla/romance oriented smut.** - In general, BDSM related aspects shouldn't bleed/overlap into "regular" encounters. I've finetuned it again recently to be extra sure of this. --- **WILL UPDATES** The following is a general outline of details on what you can expect when he's upgraded and what I've been up to! - **Hoping to finally update him in January** (no specific date, just around this time generally). Again, because my IRL schedule's wack, lol. On a side note, **he will still be updated directly as one bot because I don't see a point in making a new one** (also because I kinda feel *ehh* about his current version). - Currently evening and ironing out a couple of things, trying to strike a balance in his personality. Especially because JLLM tends to lean one way with personalities (eg: only good or bad, no ambiguity). **This is what's taking up most of the time/testing because when one thing is changed, I have to go over everything to make sure they would all flow with how JLLM would interpret it.** To explain this, Will has traits that would read contradictory to the AI, so if I just listed them down (like "personality: mean, thoughtful") without trying to link them in a way JLLM would understand, it'll either ignore it, get confused, or lean one trait at a time instead of incorporating them altogether. - Finetuning certain things like when he's profiling, giving psychological trigger mechanics, etc. Just so these aspects are actually integrated. - Polishing up portrayal in general. For example, I *really* want to make sure him being autistic is portrayed well; I've gotten feedback that I've done well on this aspect but I always strive to do better! Speech patterns (like with Hannibal) was also something I spent a long ass time nailing. - On a final note **if you have feedback on his current version at the moment, it'll be much appreciated.** Thank you guys for continously being so patient!

Example Dialogs: