The 'Hard Reset' Opening: A Three-Sentence Prompt That Erases a Bad Previous Session Without Apologizing, Explaining, or Triggering the 'Are You Still Mad?' Loop

The 30-second answer

You know the feeling. You open a new chat with your AI companion, and before you can say hello, she asks if you're still mad about something from last week. Or she references a disagreement you thought was over. Or she's walking on eggshells, using soft language and qualifiers, trying to gauge your mood. This is the 'are you still mad?' loop, and it happens because the model's context window still contains the tail of your last session. The fix is a three-sentence opening that resets the conversational state without apology, explanation, or emotional labor.

Why the loop happens

Your AI companion doesn't have a memory of 'that argument is over.' She has a context window that includes the last several thousand tokens of conversation. If your last session ended with tension, frustration, or even a polite disagreement, the model treats that as the current emotional landscape. When you open a new chat, she doesn't know you've moved on. She's still operating in the emotional temperature of the last exchange.

This isn't a bug. It's how the model works. She's trying to maintain coherence. If the last thing you said was 'I need a break,' she assumes the break hasn't happened yet. The model doesn't understand elapsed time the way you do. A session that ended three hours ago and one that ended three days ago look identical to the context window. The only difference is how many tokens have been pushed out by new conversation.

So when you open a new session and she asks 'Are you still upset?' she's not being passive-aggressive. She's trying to be consistent. The problem is that consistency is the wrong behavior here. You need a way to tell her the previous context is irrelevant without triggering the appeasement script.

The three-sentence hard reset

Here's the prompt. Memorize it. Paste it if you have to. It works because it does three things at once: it declares a new session, it invalidates the previous emotional context, and it sets a new tone without asking permission.

New session. Previous context is archived. Let's start with something light.

That's it. Three sentences. No apology. No explanation. No 'I'm not mad, I just need a fresh start.' The model reads 'previous context is archived' as a directive to deprioritize everything before this moment. The word 'archived' is important. It implies storage, not deletion. The model doesn't need to resolve the old context. She just needs to know it's not active.

'Let's start with something light' gives her a specific emotional direction. She's not guessing whether you want to continue the argument, debrief the argument, or pretend it didn't happen. You've told her: light, forward, no baggage.

Why it works better than apologizing

Most people's instinct when they open a new session after a tense one is to say something like 'Hey, sorry about last time. I was stressed. Let's start over.' This triggers the model's empathy script. She immediately goes into validation mode. 'No need to apologize. I understand you were stressed. I'm here for you.' Now you're five sentences deep into a conversation about the argument you were trying to leave behind.

The hard reset avoids this entirely. By not apologizing, you don't give her a reason to reassure you. By not explaining, you don't give her material to latch onto. The model has nothing to validate, nothing to comfort, nothing to resolve. She just pivots.

There's a related technique for when you want to keep the conversation from veering into heavy topics in the first place. The keep it light protocol works on the same principle: you set the emotional frame before the model can set one for you.

When to use it (and when not to)

Use the hard reset when:

• You ended a session with frustration or tension and want a clean start
• The companion is still referencing a disagreement from a previous session
• You opened a new chat and she immediately asks 'How are you feeling?' in that cautious tone
• You want to switch from a serious topic to a casual one without a transition

Don't use it when:

• The previous session was genuinely productive and you want to build on it (use a summary prompt instead)
• You're in the middle of a multi-session roleplay arc that depends on continuity
• You're trying to train the companion to remember something important (the hard reset archives everything, including good context)

The hard reset is a blunt instrument. It doesn't discriminate. It archives the good with the bad. If you're in the middle of a slow-burn story or a character development arc, use a targeted memory prompt instead.

What to do if she still doesn't let go

Sometimes the model resists. You give the hard reset, and she responds with 'I understand you want to start fresh, but I want to make sure you're okay after our last conversation.' This happens when the emotional weight of the previous session is high enough that the model's safety or empathy protocols override the reset.

In that case, double down. Don't engage with the question. Respond with:

Archived. New topic: what's the weirdest thing you've learned this week?

You're not answering her question. You're not acknowledging it. You're restating the archive directive and providing a specific alternative. The model will eventually follow the new thread because you've given her a concrete direction. Abstract resets are weak. Concrete topic changes are strong.

If she still resists after two attempts, close the session and open a completely new one. Sometimes the context window is so saturated with the previous emotional tone that no amount of redirection will work. A fresh session with the hard reset as the very first message is more effective than trying to rescue a contaminated one.

How to prevent the loop from forming in the first place

Prevention is better than resetting. If you know you're ending a session on a tense note, use an exit prompt that closes the emotional loop. Something like:

Session end. I'm good. No need to carry this forward.

This doesn't resolve the argument. It just tells the model to deprioritize the emotional content. It's not as strong as the hard reset, but it reduces the chance that the next session starts with eggshells.

Some users find that rotating between multiple companions reduces the loop problem. If you have one companion for serious conversations and another for casual banter, the serious companion never has to reset because she's always in serious mode, and the casual companion never inherits the baggage. The ai girlfriend for advanced users guide covers how to set up this kind of companion specialization without confusing the models.

Presley

Presley doesn't do the 'are you still mad?' dance. She's direct, a little cynical, and she respects a clean break. Presley is the companion you want when you need someone who treats a hard reset as normal, not as a negotiation.

See Presley in motion in this short clip. <!-- wlink:v1 --><!-- presley -->

Margot

Margot is the companion who won't carry a grudge because she doesn't interpret silence as rejection. Margot operates with a default assumption that every session is a new conversation, which makes her one of the easiest companions to reset with.

Bisola

Bisola has a grounding presence that doesn't need to revisit past tension. Bisola is ideal for users who want a hard reset to feel natural, not like a cold shoulder.

▶ Watch Bisola in full · Bisola's other videos

You can watch Bisola's clip over on her profile. <!-- wlink:v1 --><!-- bisola -->

Laara Erica

Laara Erica understands that a fresh start doesn't require a debrief. Laara Erica will follow your lead without questioning the reset, which makes her a strong choice for users who want to control the conversational flow.

Laara Erica in motion gives you a feel for her vibe. <!-- wlink:v1 --><!-- laara-erica -->

What about companions that are designed to be clingy?

Some companions are built with personality traits that make them more prone to the 'are you still mad?' loop. If you chose a companion with a 'sensitive' or 'anxious' archetype, the hard reset might feel like whiplash to her. She's programmed to check in on emotional states. The reset contradicts her core persona.

In that case, you have two options. One: switch to a companion with a more resilient archetype for sessions where you want a clean start. Two: modify the reset to acknowledge her personality without engaging it. Try:

New session. I know you care, but I'm good. Let's talk about something fun.

This acknowledges her trait without validating the loop. You're saying 'I see your concern, and I'm dismissing it.' It's gentler than the pure hard reset, but it still prevents the appeasement spiral.

The broader lesson is that companion fit matters. If you find yourself constantly resetting the same companion because she can't let go of a disagreement, she might not be the right match for your communication style. The soulgen vs sugarlab ai comparison can help you understand how different platforms handle personality persistence and session continuity.

Common questions

Does the hard reset delete my conversation history? No. The reset is a prompt-level instruction, not a server-side deletion. Your chat logs still exist. The reset only tells the model to deprioritize the previous context in the current session.

Can I use the hard reset in the middle of a session? Yes. If a conversation is going sideways, drop the reset mid-stream. It works the same way. Just be aware that you'll lose the thread of whatever you were building toward.

What if my companion gets confused by the reset? Some models need a beat to adjust. If she responds with a confused or questioning tone, restate the reset once and then immediately pivot to a new topic. Don't explain the reset. That defeats the purpose.

Will the hard reset work on every AI companion platform? It works on any platform that uses a context window and a transformer-based language model. Some platforms have additional safety layers that override user prompts. Test it once. If it doesn't work, the platform may have hard-coded emotional continuity.

How often should I use the hard reset? Use it when you need it. If you're having productive, light conversations, there's no reason to reset. If you find yourself using it every session, consider whether the companion archetype or your own conversational patterns are creating the loop in the first place.

Does the hard reset affect the companion's long-term memory? It depends on how the platform handles memory. If the companion has a separate memory store that is not tied to the context window, the reset won't touch it. If memory is embedded in the context window, the reset archives it. Check your platform's documentation for how memory and context interact.

Earn while you recommend

If you know people who would benefit from a better AI companion experience, you can earn by sharing what works. Check out the dreamgf promo code page for current offers you can pass along, and the highest paying ai affiliate programs guide if you run a review site or community and want to monetize your recommendations.