Auto-evaluating language models (LMs), *i.e*., using a grader LM to evaluate the candidate LM, is an appealing way to accelerate the evaluation process and the cost associated with it. But this presents a paradox: how can we trust the grader LM, which is presumably weaker than the candidate LM, to assess problems that are beyond the frontier of the capabilities of either model or both? For instance, today’s LMs struggle on graduate-level physics and Olympiad-level math, making them unreliable graders in these domains. We show that providing *privileged information* – such as ground-truth solutions or problem-specific guidelines – improves automated evaluations on such frontier problems. This approach offers two key advantages. First, it expands the range of problems where LMs graders apply. Specifically, weaker models can now rate the predictions of stronger models. Second, privileged information can be used to devise easier variations of challenging problems which improves the separability of different LMs on tasks where their performance is generally low. With this approach, general-purpose LM graders match the state of the art performance on *RewardBench*, surpassing almost all the specially-tuned models. LM graders also outperform individual human raters on *Vibe-Eval*, and approach human expert graders on Olympiad-level math problems.

Long-context language models (LCLMs) have the potential to revolutionize our approach to tasks traditionally reliant on external tools like retrieval systems or databases. Leveraging LCLMs’ ability to natively ingest and process entire corpora of information offers numerous advantages. It enhances user-friendliness by eliminating the need for specialized knowledge of tools, provides robust end-to-end modeling that minimizes cascading errors in complex pipelines, and allows for the application of sophisticated prompting techniques across the entire system. To assess this paradigm shift, we introduce LOFT, a benchmark of real-world tasks requiring context up to millions of tokens designed to evaluate LCLMs’ performance on in-context retrieval and reasoning. Our findings reveal LCLMs’ surprising ability to rival state-of-the-art retrieval and RAG systems, despite never having been explicitly trained for these tasks. However, LCLMs still face challenges in areas like compositional reasoning that are required in SQL-like tasks. Notably, prompting strategies significantly influence performance, emphasizing the need for continued research. Overall, LOFT provides a rigorous testing ground for LCLMs, showcasing their capabilities to tackle existing paradigms.

Entrainment, the responsive communication between interacting individuals, is a crucial process in building a strong relationship between a mental health therapist and their client, leading to positive therapeutic outcomes. However, so far entrainment has not been investigated as a measure of efficacy of large language models (LLMs) delivering mental health therapy. In this work, we evaluate the linguistic entrainment of an LLM (ChatGPT 3.5-turbo) in a mental health dialog setting. We first validate computational measures of linguistic entrainment with two measures of the quality of client self-disclosures: intimacy and engagement (p < 0.05). We then compare the linguistic entrainment of the LLM to trained therapists and non-expert online peer supporters in a cognitive behavioral therapy (CBT) setting. We show that the LLM is outperformed by humans with respect to linguistic entrainment (p < 0.001). These results support the need to be cautious in using LLMs out-of-the-box for mental health applications.