What to do next
You have conference + workshop access. Open your CoLM account for the badge QR at check-in. Your saved Friday agenda contains four workshops at the same time; pick which rooms to visit.
The calendar download includes four Luma events marked Going. For saved talks and workshops, open My CoLM agenda and use its Subscribe or Download .iCal link. If Luma already placed an event on your calendar, importing it again may create a duplicate.
Tuesday · Oct 6
Wednesday · Oct 7
Thursday · Oct 8
Friday · Oct 9
Four workshops are saved in your CoLM agenda, all 8:30 AM–6 PM. Choose sessions and rooms rather than trying to attend all four in full.
Pick the poster you’d rather visit
Ten AI reviewers independently swiped 30 paper pairs each. Your 20 picks then reshape the ranking. Tap a card to choose it; your choices stay on this device.
Your poster route
Eight predicted matches in each session, extrapolated from your 20 choices to all 856 accepted papers. Scores are relative, not probabilities; your direct choices are stronger evidence than predictions for unseen papers.
Download top-three calendar route from your first 20 swipes ↓
Overall ranking
Explore all accepted papers
How the ranking works
All 856 accepted abstracts are searchable. A broad interest screen selected 60 papers for 30 balanced comparisons. Ten separate reviewer agents read the paired abstracts and each made 30 independent choices. Their recorded votes form an Elo-based starting ranking. Your 20 choices train a regularized pairwise preference model on abstract text, with a small direct-choice Elo component. The model scores all papers and picks the top ones in each poster session. This is a navigation aid, not a judgment of paper quality. Check each official poster page for schedule changes.
See the reviewers’ recorded choices
People & booths to find
These are the actionable saved posts. Open a post for context before approaching someone.
Offering a few 1:1 spots; hiring PhD students in interpretability, safety, reasoning, and understanding.
See post ↗At CoLM; interested in collaborative assistants, multi-agent systems, and social intelligence.
See post ↗At CoLM; discussing AI manipulation, threat modeling, evals, and technical governance.
See post ↗Inviting conversations on synthetic data and reward models.
See profile ↗Presenting a paper on creativity in LM research agents; ask about exploration versus exploitation.
See post ↗Visit the live RSI Arena demonstration if the topic interests you.
See post ↗Workshop on long context opportunities, limits, and safety; check its social details.
See post ↗