Adaptformer: Sequence Models as Adaptive Iterative Planners
Bibliographic record
Abstract
Despite recent advances in sequence models for autonomous systems, adapting to harder, unseen tasks at test time while leveraging only demonstrations from simpler tasks remains a significant challenge. This limitation is particularly critical in planning and decision-making, where agents must utilize previously observed data to generate informed actions for novel scenarios, rather than resorting to random behavior. Conventional behavioral cloning techniques often fail in these contexts, as they rely heavily on well-represented demonstrations (labeled data) and struggle with coherent generation of long-horizon plans. To address these challenges, we propose Adaptformer, a stochastic and adaptive planner that leverages energy-based sequence models to enable sample-efficient exploration and exploitation. Adaptformer learns an energy landscape to optimize trajectories and adapts effectively to novel test cases. Additionally, it employs an intrinsic goal proposal module to generate achievable shorter sub-goals, enabling consistent and long-horizon action sequence generation. This modular design also facilitates reasoning about the planner’s behavior. Empirical evaluations in procedurally generated maze environments demonstrate Adaptformer’s effectiveness, achieving up to a 25% improvement over state-of-the-art methods in long-horizon adaptation tasks where existing models fail.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".