Skip to content

[OEL] Table 2 "In-Context": K=10 aggregation, act-budget truncation, and a possible budget-dependent reversal of the self-vs-cross ordering #441

Description

@Estrellajer

Hi, thanks for releasing the OEL code. We reproduced the Frozen Lake in-context experiments from the pinned commit (4f2a9de) and have two questions plus one observation we'd like to check with you.

  1. How is the "In-Context" column of Table 2 aggregated over the K=10 experiences?
    ray_trainer.py supports in-context evaluation via trainer.eval_prepend_experience / trainer.experience_path (one experience file per invocation), but scripts/textgame_eval.sh doesn't expose these keys, and we couldn't find the outer loop that produces the reported numbers. Is it: run the held-out 128-seed evaluation once per experience file, then average the 10 pass rates? And does "averaged over 10 random seeds" refer to the K=10 extraction seeds, or to independent evaluation seeds? (This matters more than it looks: in our runs, per-draw quality for the 1.7B self experience varies hugely — per-draw pass rates ranged 0.26–0.91, sd ≈ 0.20 — while 4B/8B experiences are stable, sd 0.03–0.06.)

  2. What is the step-level truncation rate at max_tokens=1024?
    With Qwen3-1.7B in thinking mode, we measure that most per-step responses naturally exceed 1024 tokens when generation is not budget-limited (mean ≈ 1,500 tokens; 50–74% of steps exceed 1024, depending on the experience arm). Could you share the truncation / unparseable-action rates you observed under the paper's configuration?

  3. Observation: at a non-binding budget, the self-vs-cross ordering reverses.
    Matching your protocol (official 128 held-out seeds, K=10 with seeds 50..500, val sampling 0.7/0.8/20, v3/v4 extraction prompts verbatim, sequential accumulation) but raising the per-step budget to 4096 so that >96% of actions parse, we get In-Context pass rates of: no-experience 0.48, self (1.7B) 0.60, Qwen3-4B 0.88, Qwen3-8B 0.83, larger models ≈ 0.85 — i.e., cross-model experience is decisively better than self experience, the opposite ordering of Table 2 (23.8 vs 18.0). One mechanism consistent with both results: self-extracted experience makes the model's thinking markedly shorter (mean 1,217 vs ≈1,550 tokens for cross arms), so at a 1024 budget the self arm has roughly twice the budget-survival rate of cross arms. If that's what drives the published ordering, the "off-policy knowledge is worse than self knowledge" conclusion may be specific to the 1024 budget rather than a property of the knowledge itself.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions