finding
active
finding:ds-v3-2-wins-10-7-of-mixed-gamesDS-v3.2 wins 10.7% of mixed games
Poor performance against code agents.
Source paper
extracted_from(2026) · Robert Müller · Clemens Müller
Related by similarity (8)
cosine ≥ 0.65 · no typed edgeEntities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.
- Half the games won against code agents.
- Robust performance against algorithmic baselines.
- Similarly poor against code agents.
- performance in mixed games against three code agents
- overbid rate for DeepSeek v3.2
- In the 172-game exp2 slice, G3-F has the highest LLM win rate against deterministic baselines.
- Very low win rate against code agents.
- G3-F achieved a win rate of 72.9% in the combined-comp1 98-game slice.