RLM: Fix trajectory collision

[snimu]

Description

1. Fixes interleaved prompt‑ID computation in RLMEnv by filtering to main‑trajectory steps, preventing sub‑LLM steps from corrupting main‑LLM tokenization (the root cause of /tokenize empty‑messages errors in
   training)
2. Factors model calls into Environment._call_model_api, so main and sub‑LLMs share the same request/validation path
3. Unifies sub‑LLM interleaving with the main model (incremental prompt IDs) and adds coverage in tests/test_rlm_env.py
4. RLM‑secrets dataset updates: add task/example_id, remove eval dataset/num_eval_examples, update README

Type of Change

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• Documentation update
• Test improvement

Testing

• All existing tests pass when running uv run pytest locally.
• New tests have been added to cover the changes

Checklist

• My code follows the style guidelines of this project as outlined in AGENTS.md
• I have performed a self-review of my own code
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• My changes generate no new warnings
• Any dependent changes have been merged and published

---

Note

• Shared model-call path: Factors request/validation into Environment._call_model_api, used by both main and sub-LLM paths (chat, tokens endpoint), with overlong-prompt handling.
• Interleaved tokenization fix: RLMEnv.get_model_response filters state["trajectory"] to the main trajectory_id before get_prompt_ids; sub-LLM calls keep a minimal sub_state to compute incremental prompt_ids across tool turns.
• Sub-LLM call path: Normalizes message content, supports /chat/completions/tokens with return_token_ids, and avoids direct chat.completions.create when interleaved. Adds tests for tokens endpoint usage, message normalization, and incremental prompt IDs.
• RLM-secrets updates: Dataset rows now include example_id and task; removes eval dataset/num_eval_examples. README notes eval via re-instantiating with a different seed.

^{Written by Cursor Bugbot for commit fd52bee. This will update automatically on new commits. Configure here.}

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

^{Bugbot Autofix is OFF. To automatically fix reported issues with Cloud Agents, enable Autofix in the Cursor dashboard.}

[willccbb]

hmmm would it be possible to have this handled by overriding add_trajectory_step from MultiTurnEnv? maybe we still want to log those steps, but they shouldn't be the in the main sequence in state['trajectory'] if they won't be used for training IMO

this feels like some of the RLM logic is creeping a bit too low into the stack (base Environment shouldn't know what an RLM is or have to think about it), and preparing the context for the API call should probably be handled by get_prompt_messages. If the RLM environment promises that get_prompt_messages only ever contains "increasing" sequences of messages on the subset of turns where a step will be added to state['trajectory'], then the old/current approach should work without changes I think?

In general though, I'm a bit skeptical about trying to shoehorn RLMs into the interleaved strategy. Ultimately we want a single "best of both worlds" strategy which is always TITO, always "just works" with get_prompt_messages, and aggressively interleaves until a message sequence forces a branch