Christoph Riedl
@criedl
Followers
1K
Following
311
Media
58
Statuses
2K
Professor for Information Systems, Northeastern University; Interested in collective intelligence, human-AI teaming & crowdsourcing
Boston, MA
Joined January 2010
Evidence that multi-agent LLMs develop emergent coordination and specialized roles—especially with personas and "think about others" prompts. No free lunch tho: strong performance needs both alignment and complementarity https://t.co/YM8McXGhaY
0
5
10
How does behavior spread? Do interactions beyond pairwise ties matter? We build on hypergraphs to explore higher-order social influence and study a country-scale randomized experiment
0
0
1
Complementary read to our other recent paper on complex contagion https://t.co/W8CyrS8M6s
pnas.org
How does social network structure amplify or stifle behavior diffusion? Existing theory suggests that when social reinforcement makes the adoption ...
0
0
0
Across three different analytical methods we provide evidence for complex contagion: the contagion process cannot be understood as independent cascades but rather as a process in which signals from multiple sources amplify each other through synergistic interdependence
1
0
0
How does behavior spread? New insights on complex contagion in social networks with causal evidence from a country-scale field experiment @davidlazer @jaemin_lee_phd
https://t.co/UaXknmxz7N
0
2
6
This is wild 🤯 Multi-agent LLMs are starting to think together. A new paper "Emergent Coordination in Multi-Agent Language Models” just dropped, and it’s mind-blowing. Researchers found that when you connect multiple LLMs with light feedback (and no direct communication),
41
153
666
A lot of problems with AI discourse are because "being good at AI" (called Theory of Mind in this paper) is a skill that seems to be independent of "being great at your job" So you have amazing experts who gain from AI, and others who do not, and they don't understand each other
45
101
773
The #1 Skill to Master AI Isn't What You Think (Hint: It's Not Prompt Engineering). 1/16 You know that frustrating feeling when you ask an AI a question and it gives you a useless answer? You tweak the prompt, add more detail... and it's still not quite right. What if the
9
37
163
Emergent properties have functional benefits and realign internal group structure. Results mirror collective intelligence principles: effective performance requires both alignment on shared objectives and complementary contributions across members.
1
1
0
We find multi-agent systems have capacity for emergence - they are real "teams" that are more than the sum of their parts. And we can steer them with clever prompts. The ToM condition in particular leads to stable specialization and goal-directed complementarity across agents
1
1
1
Experiments use a simple guessing game without direct agent communication and only minimal group-level feedback with three randomized interventions: plain, agents with personas, and personas with an instruction to "think about what other agents might do" (a ToM prompt)
1
1
0
When are multi-agent LLM systems merely a collection of individual agents versus an integrated collective with higher-order structure? New paper shows multi-agent LLM systems have capacity for emergent coordination and how to steer them ...
1
1
3
📌Save the Date! The flagship conference of the Network Science Society - 𝗡𝗲𝘁𝗦𝗰𝗶 𝟮𝟬𝟮𝟲 - is coming to Northeastern University’s Network Science Institute, 𝗝𝘂𝗻𝗲 𝟭-𝟱, 𝟮𝟬𝟮𝟲. Registration opens soon! 🔗 https://t.co/LH8ikP4bTj
0
17
28
Who is going to be at #COLM2025? I want to draw your attention to a COLM paper by my student @sheridan_feucht that has totally changed the way I think and teach about LLM representations. The work is worth knowing. And you meet Sheridan at COLM, Oct 7!
[📄] Are LLMs mindless token-shifters, or do they build meaningful representations of language? We study how LLMs copy text in-context, and physically separate out two types of induction heads: token heads, which copy literal tokens, and concept heads, which copy word meanings.
3
38
188
Some useful findings: 1) Working with AI boosts the performance of people solving math, science & ethics questions 2) The biggest boost is for the hardest problems 3) High performers remain highest performing, but low performers gain more 4) People who are good with AI gain most
16
77
508
New paper on quantifying human-AI synergy in a large dataset with AI-alone, human-alone, and human-AI together. And
1
1
4