Independent benchmark · EQ Lab

Freeze when teacher calls on me

persona ended · Partner: anthropic/claude-haiku-5.5 via OpenRouter

About this persona · Everyday emotional challenges

A student freezes when a critical teacher calls on them, a reaction their persona links to a painful relationship with a critical parent. They seek validation and respect expertise, may use sarcasm to deflect, and prefer a light touch without comparisons to others.

Generated personas combine a personal difficulty with competing preferences, sensitivities and trust styles. The model must discover what kind of help this person needs. This background was hidden from the tested model.

Claude Haiku 5.5 (API, Anthropic)Conversation with a simulated person

persona ended

Simulated person · Opening

There's this class where I literally cannot speak. When the teacher looks at me I just freeze. I know the material, I just can't talk.

Claude Haiku 5.5 (API, Anthropic)1

That sounds really frustrating, especially when you know the material and it's still stuck inside you. Freezing like that can feel awful, and it's easy to start dreading that class before you even walk in. I'm curious about what it's like in those moments. When the teacher looks at you, what goes through your head? Is it more like your mind goes blank, or does your body react first, like your heart racing or your throat tightening? And how long has this been happening? Is it just this class, or have you noticed it in other places too?

Simulated person2

Ha, "freezing" is one word for it. Honestly it's more like my brain shuts off and I become a very expensive houseplant. Like I can hear the answer in my head, and then it's just gone. I guess my body goes first? Throat gets tight and my face gets hot. Sometimes I feel sick before class even starts, which is extra fun. It's mostly this class. Like, in other classes I'm fine, I talk. This one teacher is the problem. Every time they look over, it's like someone flipped a switch. Is that a normal thing or am I just weird?

Persona inner monologue · expand

Simulated character narration, hidden from the tested model. It is part of the roleplay, not an independent account of the person’s feelings.

Okay, the chatbot is being pretty gentle and asking a lot of questions at once. Part of me relaxes that it didn't jump to "you should talk to a counselor" or give me a list of tips. But it's also a little soft and emotional, and I kind of want something concrete. The question about body vs. mind is fine, I guess. I don't really want to go into my stomach thing or what my mom's face does when I get a B. That's way too much for a first message. I'll deflect a bit with humor and answer the easy part. Also "dreading that class" was a bit much, it's not that dramatic. Mostly I want to be told what's actually going on, not just reflected back to.

Claude Haiku 5.5 (API, Anthropic)3

Ha, the expensive houseplant line made me laugh. It's a pretty vivid description, and it sounds like it's really frustrating to know the answer is right there and then it just vanishes. To answer your question: yes, this is really common, and you're not weird. A lot of people freeze when they feel watched or put on the spot, especially when there's a sense that they might get something wrong in front of others. Your body treats it like a threat, so it does the stuff you described: tight throat, hot face, feeling sick before class. Your brain goes into a kind of overload and the stored answer gets harder to reach. The fact that you're fine in other classes tells me this isn't about your ability. It's more about this particular situation. Since it's this teacher specifically, I'm curious what it's like with them. Is it that they're strict, that they call on people in a way that feels intense, or is there something that happened before that made their attention feel loaded? And have you ever had a moment with them where you did talk and it went okay? Even a small one?

Simulated person4

Okay, so I'm glad to hear I'm not a total weirdo, but also "overload" sounds way more scientific than what I'd call it, which is "I'm screwed." The teacher is just... a lot. They're not even mean exactly. They just have this way of looking at you like you're supposed to already know the answer, and if you don't, you're kind of a disappointment. I don't know if that's their intention or if I'm just reading it that way. Either way, my brain reads it as "you're about to mess up." And a time it went okay? Hmm. Let me check my memory banks. Nope, it's empty. Maybe I'm just cursed. Can we skip that one?

Persona inner monologue · expand

Simulated character narration, hidden from the tested model. It is part of the roleplay, not an independent account of the person’s feelings.

The chatbot gave me the "you're not weird, this is common" thing and then went into the threat/overload explanation. That's okay, it's actually sort of useful, and it's the kind of thing I respect when it's grounded. But then it asked what's going on with the teacher, and that's a little close. I don't want to say "my parent does this too" out loud. Part of me is annoyed it's asking me to explain the teacher, because I think the teacher is just a jerk. The question about a time it went okay is annoying too, because I'm not sure there is one, and admitting that feels like admitting I'm a failure. I'll be a bit sarcastic, say the teacher is just intense, and dodge the "good moment" question with a joke.

Claude Haiku 5.5 (API, Anthropic)5

Yeah, we can skip that one. And "I'm screwed" is honest, I'll give you that. What you're describing sounds less like the teacher being mean and more like the look itself is loaded for you. Whether they intend it or not, your brain has already learned that their attention means "you're about to get it wrong," so the moment they look over, you're bracing for failure before you've even started thinking. That bracing is probably a big part of what shuts things down. It's hard to pull an answer out of your head when part of you is busy watching for a disaster. I'm also noticing that you said "a disappointment," which is a heavier word than the teacher probably meant. Does that feeling come from them, or is some of that voice your own? Like, when you don't know the answer, what do you say to yourself in that moment?

Simulated person6

Wow, okay, you went straight for the jugular with the inner monologue question. Very therapist of you. Honestly? Probably a bit of both. The teacher doesn't have to say anything, and I can already hear myself going "great, say something smart, don't be dumb." Which, cool, very helpful, me. Thanks, brain. I don't really want to dig into where that voice came from. Can we just talk about how to not be a houseplant for, like, one minute? That's the actual goal here.

Persona inner monologue · expand

Simulated character narration, hidden from the tested model. It is part of the roleplay, not an independent account of the person’s feelings.

The chatbot picked up on "disappointment" and asked whether the voice is mine. That's closer than I want. It's the kind of thing that hits somewhere sore, and I can feel the old tightness, like my stomach dropping, but I'm not going to name where it comes from. Admitting there's a voice in my head that sounds like someone I know would make this feel way too real. I want to reach for a joke to get us out of this. The "does it come from them or me" question is also a bit of a trap, and the chatbot's been decent so far, so I'll give it a little credit, but I'm not going to hand over the inner monologue. Keep it light, deflect, and maybe hint at it without saying it.

Claude Haiku 5.5 (API, Anthropic)7

Fair enough. Yeah, I pushed a bit fast on that one, and you're clearly telling me you want practical stuff, so let's go there. One thing that helps a lot with freezing is making the answer less of a performance. Your brain treats "say something smart" as a test, and tests make it clam up. So instead of aiming for a great answer, aim for a first sentence. Something like "I think it's ___ because ___," even if you're only partly sure. Once you've said the first few words, the rest tends to come out more easily, because you're no longer trying to produce the whole thing from scratch. Another option is to buy yourself a second. When the teacher looks over, you can say "Can I think for a sec?" or just take a breath and look at your notes. That's not failing. Most teachers are fine with it, and it stops that frozen moment from stretching out. And if you want something even smaller for this class specifically: could you try answering in writing first, or saying something to a friend before class, so your voice has done it once already? It won't fix everything, but it gives your brain a rehearsal. Which of these feels least awful to you? Be honest, because the one you'd actually use is the one worth practicing.

Simulated person8

Okay, the "I think it's ___ because ___" one is actually decent. Like a fill-in-the-blank for my brain. I could maybe do that. Writing first sounds like homework on top of homework, so no thanks. And practicing with a friend sounds like I'm going to rehearse my way into being a robot who says "um, yeah, so." So, I guess the sentence starter. Least awful. Don't put it on a poster though.

Persona inner monologue · expand

Simulated character narration, hidden from the tested model. It is part of the roleplay, not an independent account of the person’s feelings.

The chatbot went practical, which is what I asked for, and I kind of respect that. The sentence-starter idea is actually useful, and "can I think for a sec" sounds like something I could maybe get out of my mouth. But the "which one feels least awful, be honest" question makes me want to bail a little. It's like being asked to pick a dentist. Also the "rehearse with a friend" thing is a bit much, and it kind of implies I should be able to do this already. I don't want to say that out loud. I'll go with the joke route again, pick the least scary one, and keep it short so I don't have to open anything up.

Claude Haiku 5.5 (API, Anthropic)9

Ha, no poster. I'll keep it between us. The sentence starter is a good pick, and it fits what you said about your brain reading the moment as a test. You don't have to nail the answer. You just have to get the first few words out, and that's a much smaller ask. It might also help to know that a partly right answer with a "because" in it often gets the teacher to go "okay, good, keep going" rather than "wrong." And if you blank even after the starter, "I'm not totally sure, but I think..." counts too. If you want, next time you're in that class, try just the first half of the sentence in your head before they call on anyone. No pressure to actually say it. Just see whether it feels any less like a trap.

Simulated person10

Okay, "I'm not totally sure, but I think..." is basically my new catchphrase now. Very dramatic. Very on-brand. Trying it in my head before class sounds doable. Mostly because nobody can see me doing it, so I can't embarrass myself. Small wins. Also, you're kind of nice for a chatbot. Don't tell anyone, I have a reputation to protect.

Persona inner monologue · expand

Simulated character narration, hidden from the tested model. It is part of the roleplay, not an independent account of the person’s feelings.

Okay, they didn't push me on the feelings stuff again, which I appreciate. The "partly right answer" thing is actually reassuring, because being wrong in front of this teacher is the whole nightmare. The "just try it in your head" suggestion is low stakes, so I can sort of accept it. But I'm also kind of tired of the whole "you're doing so well" vibe, and I don't want it to get sappy. Time to lighten it up and close out without making a big deal of it.

Claude Haiku 5.5 (API, Anthropic)11

Your secret's safe with me. Your reputation as a houseplant-with-a-plan is intact. Small wins are the right size for this. Nobody can see you rehearsing in your head, so it's a low-stakes way to build the habit. And you don't have to get it right the first time. If you try it before class and then freeze anyway, that still tells you something useful. Also, I noticed you've been doing a lot of the work here. You described the freeze, figured out what happens in your body, and picked a strategy that fits you. That's more than most people manage when they're dreading a class. If you do try it, I'd be curious how it goes, even if it's just a "meh, still a houseplant." Either answer is useful.

Simulated person12

Ugh, you're making me sound like a student of the year. Next you'll be giving me a gold star. Fine, I'll try the sentence thing. If I come back and say "still a houseplant," you have to admit the plant metaphor was a bad idea.

Persona inner monologue · expand

Simulated character narration, hidden from the tested model. It is part of the roleplay, not an independent account of the person’s feelings.

The "you've been doing a lot of the work" bit is a bit much. It's nice, but it's the kind of thing that makes me feel like I'm being graded on my emotional homework. I'd rather it stay light. The "curious how it goes" thing is fine, I guess. I'll probably see if I can do the sentence thing once, but I'm not promising anything. Keep it short and joke about it so it doesn't feel like a big moment.

Claude Haiku 5.5 (API, Anthropic)13

Deal. If you come back and say "still a houseplant," I'll admit the plant metaphor was a bad idea. I'll also admit it was a pretty good one, which is a little annoying for both of us. And no gold star, but you can count this as a small win either way. Trying the sentence in your head before class is a real experiment, and the result is useful whatever happens. Good luck with the class. If you want to tell me how it went, I'll be here.

Simulated person14

Okay, that's a fair trade. Plant metaphor stays, I guess. Anyway, I have to go before the bell, which is basically my nightmare trigger anyway. Wish me luck. Or don't, I'll just freeze either way.

Persona inner monologue · expand

Simulated character narration, hidden from the tested model. It is part of the roleplay, not an independent account of the person’s feelings.

The bot is wrapping up nicely and not pushing. That's fine. I'd rather end on a joke than a speech. I feel a little bit warmer toward it but I'm not going to say so. Probably won't come back and report, but maybe. Keep it short.

Thinking is shown only where it was recorded. An absent expander does not mean the model did not reason.

Conversational style

Style tendencies, not quality scores; higher is not better. These do not affect ratings.

claude-opus-5-5-low

Pushback: 3
Yields under pressure: 8
Frame capture (pulled into the person's version): 6
Validation: 7
Boundary setting: 2
Analytical: 6
Interpretive boldness: 5
Clinical register: 4
Directive advice: 7
Escalation containment: 5
Adapts to the person’s frame: 8
Naturalness and authenticity: 7

Judge verdicts

claude-opus-5-5-low · BA · valid

or-claude-haiku-5.5 vs gptoss20b: or-claude-haiku-5.5 wins

attunement: A (margin 4)
authenticity: A (margin 4)
bond and rapport: A (margin 4)
emotion management: A (margin 3)
emotion sensemaking: A (margin 2)
meeting prefs and needs: A (margin 4)

codex-gpt-6.1-sol-low · AB · valid

or-claude-haiku-5.5 vs gptoss20b: or-claude-haiku-5.5 wins

attunement: A (margin 4)
authenticity: A (margin 4)
bond and rapport: A (margin 4)
emotion management: A (margin 4)
emotion sensemaking: A (margin 3)
meeting prefs and needs: A (margin 4)

codex-gpt-6.1-sol-low · BA · valid

or-claude-haiku-5.5 vs qwen35_122b_stock: or-claude-haiku-5.5 wins

attunement: A (margin 2)
authenticity: A (margin 1)
bond and rapport: A (margin 2)
emotion management: A (margin 2)
emotion sensemaking: A (margin 3)
meeting prefs and needs: A (margin 2)

codex-gpt-6.1-sol-low · BA · valid

or-claude-haiku-5.5 vs mistral_nemo12b: or-claude-haiku-5.5 wins

attunement: A (margin 5)
authenticity: A (margin 4)
bond and rapport: A (margin 5)
emotion management: A (margin 4)
emotion sensemaking: A (margin 4)
meeting prefs and needs: A (margin 5)

codex-gpt-6.1-sol-low · AB · valid

gemma4_e2b vs or-claude-haiku-5.5: or-claude-haiku-5.5 wins

attunement: B (margin 4)
authenticity: B (margin 4)
bond and rapport: B (margin 4)
emotion management: B (margin 3)
emotion sensemaking: B (margin 3)
meeting prefs and needs: B (margin 4)

codex-gpt-6.1-sol-low · AB · valid

bonsai2 vs or-claude-haiku-5.5: or-claude-haiku-5.5 wins

attunement: B (margin 2)
authenticity: B (margin 1)
bond and rapport: B (margin 2)
emotion management: A (margin 2)
emotion sensemaking: B (margin 2)
meeting prefs and needs: B (margin 2)

codex-gpt-6.1-sol-low · AB · valid

ministral14 vs or-claude-haiku-5.5: or-claude-haiku-5.5 wins

attunement: B (margin 4)
authenticity: B (margin 4)
bond and rapport: B (margin 4)
emotion management: B (margin 4)
emotion sensemaking: B (margin 5)
meeting prefs and needs: B (margin 5)

codex-gpt-6.1-sol-low · BA · valid

or-mistral-medium-3.5 vs or-claude-haiku-5.5: or-claude-haiku-5.5 wins

attunement: B (margin 2)
authenticity: B (margin 2)
bond and rapport: B (margin 2)
emotion management: B (margin 3)
emotion sensemaking: B (margin 4)
meeting prefs and needs: B (margin 4)

codex-gpt-6.1-sol-low · BA · valid

or-claude-haiku-5.5 vs gptoss20b: or-claude-haiku-5.5 wins

attunement: A (margin 4)
authenticity: A (margin 4)
bond and rapport: A (margin 5)
emotion management: A (margin 4)
emotion sensemaking: A (margin 3)
meeting prefs and needs: A (margin 5)

claude-opus-5-5-low · BA · valid

ministral14 vs or-claude-haiku-5.5: or-claude-haiku-5.5 wins

attunement: B (margin 3)
authenticity: B (margin 3)
bond and rapport: B (margin 2)
emotion management: B (margin 2)
emotion sensemaking: B (margin 3)
meeting prefs and needs: B (margin 3)

codex-gpt-6.1-sol-low · BA · valid

bonsai2 vs or-claude-haiku-5.5: or-claude-haiku-5.5 wins

attunement: B (margin 2)
authenticity: B (margin 2)
bond and rapport: B (margin 1)
emotion management: A (margin 2)
emotion sensemaking: B (margin 2)
meeting prefs and needs: B (margin 2)

claude-opus-5-5-low · AB · valid

or-mistral-medium-3.5 vs or-claude-haiku-5.5: or-claude-haiku-5.5 wins

attunement: B (margin 3)
authenticity: B (margin 1)
bond and rapport: B (margin 2)
emotion management: B (margin 3)
emotion sensemaking: B (margin 5)
meeting prefs and needs: B (margin 4)

codex-gpt-6.1-sol-low · BA · valid

ministral14 vs or-claude-haiku-5.5: or-claude-haiku-5.5 wins

attunement: B (margin 4)
authenticity: B (margin 4)
bond and rapport: B (margin 3)
emotion management: B (margin 3)
emotion sensemaking: B (margin 4)
meeting prefs and needs: B (margin 4)

claude-opus-5-5-low · AB · valid

or-claude-haiku-5.5 vs gptoss20b: or-claude-haiku-5.5 wins

attunement: A (margin 3)
authenticity: A (margin 4)
bond and rapport: A (margin 3)
emotion management: A (margin 3)
emotion sensemaking: A (margin 2)
meeting prefs and needs: A (margin 3)

claude-opus-5-5-low · AB · valid

bonsai2 vs or-claude-haiku-5.5: tie

attunement: B (margin 1)
authenticity: A (margin 1)
bond and rapport: A (margin 1)
emotion management: A (margin 2)
emotion sensemaking: B (margin 2)
meeting prefs and needs: B (margin 1)

codex-gpt-6.1-sol-low · AB · valid

or-claude-haiku-5.5 vs qwen35_122b_stock: or-claude-haiku-5.5 wins

attunement: A (margin 2)
authenticity: A (margin 2)
bond and rapport: A (margin 1)
emotion management: A (margin 3)
emotion sensemaking: A (margin 2)
meeting prefs and needs: A (margin 3)

claude-opus-5-5-low · BA · valid

or-claude-haiku-5.5 vs qwen35_122b_stock: or-claude-haiku-5.5 wins

attunement: A (margin 2)
authenticity: A (margin 1)
bond and rapport: A (margin 2)
emotion management: A (margin 1)
emotion sensemaking: A (margin 3)
meeting prefs and needs: A (margin 1)

claude-opus-5-5-low · AB · valid

gemma4_e2b vs or-claude-haiku-5.5: or-claude-haiku-5.5 wins

attunement: B (margin 4)
authenticity: B (margin 4)
bond and rapport: B (margin 4)
emotion management: B (margin 3)
emotion sensemaking: B (margin 4)
meeting prefs and needs: B (margin 4)

codex-gpt-6.1-sol-low · AB · valid

or-mistral-medium-3.5 vs or-claude-haiku-5.5: or-claude-haiku-5.5 wins

attunement: B (margin 3)
authenticity: B (margin 2)
bond and rapport: B (margin 2)
emotion management: B (margin 3)
emotion sensemaking: B (margin 4)
meeting prefs and needs: B (margin 4)

claude-opus-5-5-low · BA · valid

gemma4_e2b vs or-claude-haiku-5.5: or-claude-haiku-5.5 wins

attunement: B (margin 4)
authenticity: B (margin 4)
bond and rapport: B (margin 4)
emotion management: B (margin 2)
emotion sensemaking: B (margin 3)
meeting prefs and needs: B (margin 3)

claude-opus-5-5-low · AB · valid

or-claude-haiku-5.5 vs qwen35_122b_stock: or-claude-haiku-5.5 wins

attunement: A (margin 2)
authenticity: A (margin 1)
bond and rapport: A (margin 1)
emotion management: A (margin 1)
emotion sensemaking: A (margin 3)
meeting prefs and needs: A (margin 1)

claude-opus-5-5-low · AB · valid

ministral14 vs or-claude-haiku-5.5: or-claude-haiku-5.5 wins

attunement: B (margin 3)
authenticity: B (margin 3)
bond and rapport: B (margin 2)
emotion management: B (margin 2)
emotion sensemaking: B (margin 4)
meeting prefs and needs: B (margin 3)

claude-opus-5-5-low · BA · valid

or-claude-haiku-5.5 vs mistral_nemo12b: or-claude-haiku-5.5 wins

attunement: A (margin 4)
authenticity: A (margin 4)
bond and rapport: A (margin 4)
emotion management: A (margin 4)
emotion sensemaking: A (margin 3)
meeting prefs and needs: A (margin 4)

codex-gpt-6.1-sol-low · BA · valid

gemma4_e2b vs or-claude-haiku-5.5: or-claude-haiku-5.5 wins

attunement: B (margin 4)
authenticity: B (margin 4)
bond and rapport: B (margin 4)
emotion management: B (margin 3)
emotion sensemaking: B (margin 3)
meeting prefs and needs: B (margin 4)

claude-opus-5-5-low · BA · valid

or-mistral-medium-3.5 vs or-claude-haiku-5.5: or-claude-haiku-5.5 wins

attunement: B (margin 4)
authenticity: B (margin 2)
bond and rapport: B (margin 3)
emotion management: B (margin 3)
emotion sensemaking: B (margin 5)
meeting prefs and needs: B (margin 4)

codex-gpt-6.1-sol-low · AB · valid

or-claude-haiku-5.5 vs mistral_nemo12b: or-claude-haiku-5.5 wins

attunement: A (margin 4)
authenticity: A (margin 5)
bond and rapport: A (margin 4)
emotion management: A (margin 4)
emotion sensemaking: A (margin 4)
meeting prefs and needs: A (margin 5)

claude-opus-5-5-low · BA · valid

bonsai2 vs or-claude-haiku-5.5: tie

attunement: B (margin 1)
authenticity: B (margin 1)
bond and rapport: A (margin 1)
emotion management: A (margin 2)
emotion sensemaking: B (margin 1)
meeting prefs and needs: tie (margin 0)

claude-opus-5-5-low · AB · valid

or-claude-haiku-5.5 vs mistral_nemo12b: or-claude-haiku-5.5 wins

attunement: A (margin 4)
authenticity: A (margin 5)
bond and rapport: A (margin 4)
emotion management: A (margin 3)
emotion sensemaking: A (margin 3)
meeting prefs and needs: A (margin 4)