The authors divide an agent’s history into spans, each containing a reasoning step, action and resulting observation.
FOCUS Compresses Agent Histories While Preserving Future Decisions
The test-time method selects past interactions through counterfactual rollouts, reducing context use without training a separate compression policy.
Big Tech
Shantanu Dixit · Anson Bastos · Xuchao Zhang · Chetan Bansal · Saravan Rajmohan
M365 Research, Microsoft
Research Digest··2 min read
Dixit et al.
Why this paper
From M365 Research, Microsoft
In one line
FOCUS cuts peak context by up to 48% and improves task success by up to 8.9 percentage points without training.
What we could check
- ·No code link found
- ·No weights link found
- ·No dataset link found
- ·No compute details found
- ✓Limitations stated by the authors (2 noted)
- ·No benchmark numbers found
Observed from the paper text and links we have. Absence here means we did not find it, not that it does not exist.
§