Sebastian Olschewski (@Seb_Olschewski) header
Sebastian Olschewski (@Seb_Olschewski) profile picture

Postdoc at Uni Basel and Warwick Business School. Interested in economic/financial/social decision-making, cognitive processes, & computational models

Joined January 2019
Photo by Sebastian Olschewski (@Seb_Olschewski)Photo by Sebastian Olschewski (@Seb_Olschewski)Photo by Sebastian Olschewski (@Seb_Olschewski)
New publication @PNASNews: Frequent winners explain apparent skewness preferences in experience-based decisions pnas.org/doi/epdf/10.1073/pn… With @Mikhail_Spektor and @gael_lemens Thread(1/7)>>>
1
14
2
38
4,726
Happy to share my new paper “Do you remember? Prosocial behavior declines in memory-based decisions” is out. We often make decisions that affect others without all the relevant information at hand. Instead, we rely on memory.
1
1
44
But does relying on memory affect how much we consider others? Across three experiments, I found that people made less prosocial choices when they had to recall consequences for others than when that information was directly available.
1
20
The findings are consistent with motivated forgetting: People may be less likely to recall information about others’ outcomes when doing so would make selfish choices harder to justify. 🔗 doi.org/10.1016/j.jebo.2026.…
19
Sebastian Olschewski retweeted
Humans are more creative than LLMs on open tasks requiring imagination. A new fascinating paper compares humans with three LLMs on two creative tasks. The closed task: write a story incorporating ten specified words, such as “house”, “cow” and “planet”. The open task: imagine you could invent anything simply by thinking about it. What would you create, and why? Creativity was rated from 1 to 10 by human evaluators who did not know whether responses came from humans or AI. The results are super interesting: In the closed task, LLMs outperform humans on average. By contrast, in the open task, humans outperform all three LLMs. Why? To provide a first potential answer, the authors model imagination as the human capacity to generate ideas beyond what instructions prescribe. Then they estimate its contribution from differences between human and AI scores. Their model attributes between 22% and 45% of humans’ creativity scores in the open task to imagination. Humans may be more creative than LLMs thanks to their capacity to jump beyond what is prescribed. * Paper in the first reply.
7
14
2
55
3,403
Sebastian Olschewski retweeted
📢 New publication! 📢 Excited to share our new paper "Preserved higher-order mentalising during strategic interactions after medial prefrontal cortex damage" in collaboration with @thepsychologist's lab - out now in Nature Communications! nature.com/articles/s41467-0…
🎉 Our paper is now published in @NatureComms Excited to share this work with my co-first author @ArkadyKonovalov and all our co-authors. 🧠 Preserved higher-order mentalising during strategic interactions after medial prefrontal cortex damage 🔗 doi.org/10.1038/s41467-026-7… (1/4)
2
6
226
Sebastian Olschewski retweeted
What is a minimal set of assumptions that can explain the majority of human behavior? We investigated this question with large-scale cognitive models🚀 CORE writes prediction errors in hierarchical episodic memories with context controling encoding, retrieval, and forgetting.
4
21
1
165
10,587
Sebastian Olschewski retweeted
Open access paper published in Cognitive Psychology: "Compressed representations and attentional competition in numeric integration for average estimations" w/ Alice Mason from Uni Bath & @Seb_Olschewski
6
1
1
61
Sebastian Olschewski retweeted
New short-form preprint in which we use Centaur to identify gaps in interpretable cognitive models and revise them accordingly using Qwen3 -- fully automated and without a human-in-the-loop. arxiv.org/abs/2505.17661
3
1
23
2,137
Sebastian Olschewski retweeted
Completely nonsensical politicization of scientistics identifying and correcting problems in science. - Gino's research was not "promoting progressive ideas" - The problems were not found by "bloggers," but by leading researchers/faculty in the same scientific discipline.
Harvard expelled a professor for faking data to promote progressive ideas...after paying her $1 million a year! The fake data was caught by bloggers, not peer review, because Libs wanted to believe the lie...and this isn't the first time! My latest in @NRO Link in comments
17
74
3
920
82,691
Sebastian Olschewski retweeted
Excited to share a new Trends in Cognitive Sciences paper that I had the pleasure to be a part of. This is an interdisciplinary perspective on the dynamics of cognitive costs, namely when these costs occur and how they impact our decisions. #neuroeconomics sciencedirect.com/science/ar…
1
3
17
622
Sebastian Olschewski retweeted
New preprint out with an amazing 40-person team! We find that Claude 3.5 Sonnet outperforms incentivised human persuaders in a >1000-participant live quiz-chat in deceptive and truthful directions!
4
27
16
147
23,466
Announcing the summer school on "Cognitive Foundations of Decision-Making" to be held from June 30 to July 4, 2025, in Ghent, Belgium. For information and the application pprocedures, see rislab.net/summer-school-202…
1
4
1
7
687
Sebastian Olschewski retweeted
We are looking for two PhD students at our institute in Munich. Both postions are open-topic, so anything between cognitive science and machine learning is possible. More information: hcai-munich.com/PhDHCAI.pdf Feel free to share broadly!
2
73
3
205
40,502
However, we argued that such a mechanism would be inconsistent with other experimental findings from our paper “Frequent winners explain apparent skewness preferences in experience-based decisions”.
1
76
We concluded that keeping track of the number of ordinal wins over all outcome pairs of two sequences remains an important behavioral mechanism in decisions from experience. But we are looking forward to future work to better understand how people process skewed distributions.
1
78
Sebastian Olschewski retweeted
🚨Out today in PNAS @PNASNews🚨 pnas.org/doi/10.1073/pnas.24… Why do people overestimate the size of politically relevant groups (immigrant, LGBTQ, Jewish) and quantities (% of budget spent on foreign aid, % of refugees that are criminals)? We analyze 100k estimates to find out🧵👇
8
95
9
349
53,422