[train][TBench][MiniSwe] Fix custom generator loss masking

[CharlieFRuan]

Before this PR, in TerminalBenchGenerator and MiniSweAgentGenerator, we naively generate loss mask for a chat history by doing, all zeros for user messages, and all ones for assistant messages.

However, this is incorrect. For instance, the generation prompt of assistant should be masked with zero, and the token \n that comes after the EOS for Qwen models should be masked with zero.

This PR fixes this by implementing get_response_ids_and_loss_mask_from_messages(), which uses the fixed-base helper encode_messages_subset(), to convert a chat history (excluding the initial prompt, which may differ by models) into response IDs, loss mask, and optionally rollout logprobs.

Besides, TerminalBenchGenerator had a bug of

prompt_ids = self.tokenizer.apply_chat_template(
    prompt,
    add_generation_prompt=True,
...

which is incorrect because the messages below will add the generation prompt.

We also add extensive unit tests. The unit tests are implemented by Opus 4.5, which are reviewed and iterated by me.

[gemini-code-assist]

Code Review

This pull request effectively refactors the loss mask generation by introducing a new utility function, get_response_ids_and_loss_mask_from_messages. This new function correctly handles the complexities of masking for different parts of a conversation, such as generation prompts and tokens following an EOS token, which is a significant improvement over the previous naive implementation. The associated bug fix in TerminalBenchGenerator regarding add_generation_prompt is also correctly addressed. Furthermore, the inclusion of extensive and thorough unit tests for the new utility function is commendable and greatly enhances the reliability of this change. I have one piece of feedback regarding a misleading error message that could hinder future debugging efforts.