Grammars of Formal Uncertainty: When to Trust LLMs in Automated Reasoning Tasks · HackerPapers discussions

Grammars of Formal Uncertainty: When to Trust LLMs in Automated Reasoning Tasks — mentioned by barthelomew

Sat, 04 Oct 2025 21:16:44 GMT

Yep! Datalog syntax for Z3 is pretty neat! We used SMT [1] in our grammars paper because it allowed the most interoperability with solvers, but our technique also works with PROLOG; as tested our at the behest of reviewers at NeurIPS. I would assume that this should also work with datalog [2]. [1] https://arxiv.org/abs/2505.20047 [2] https://github.com/antlr/grammars-v4/blob/master/datalog/dat...

Paper and discussion history. A mention is not an endorsement.

Grammars of Formal Uncertainty: When to Trust LLMs in Automated Reasoning Tasks — mentioned by barthelomew

Sat, 04 Oct 2025 20:08:21 GMT

Yep, this is a genuine problem, and this is what we term as the autoformalization gap in our follow up paper. (https://arxiv.org/abs/2505.20047) Some LLMs are more consistent between text and SMT, while others are not. (Tab 1, Fig 14,15) You can do uncertainty quantification with selective verification to reduce the "risk", for e.g. shown as the Area Under the Risk Coverage Curve in Tab 4.

Paper and discussion history. A mention is not an endorsement.

Grammars of Formal Uncertainty: When to Trust LLMs in Automated Reasoning Tasks — mentioned by barthelomew

Sat, 04 Oct 2025 20:02:29 GMT

Yep. The paper was written last year with GPT-4o. Things have become a lot better since then with newer models. E.g. https://arxiv.org/pdf/2505.20047 Tab 1, we compare the performance on text-only vs SMT-only. o3-mini does pretty well at mirroring its text reasoning in its SMT, vs Gemini Flash 2.0. Illustration of this can be seen in Fig 14, 15 on Page 29. In commercially available products like AWS Automated Reasoning Checks, you build a model from your domain (e.g. from a PDF policy document), cross verify it for correctness, and during answer generation, you only cross check whether your…

Paper and discussion history. A mention is not an endorsement.

Grammars of Formal Uncertainty: When to Trust LLMs in Automated Reasoning Tasks — mentioned by barthelomew

Sat, 04 Oct 2025 19:48:33 GMT

Hey, yes! This is because the DSL (Domain Specific Language) is pretty complex, and the LLM finds it hard. We prototype a much more effective version using SMT in our NeurIPS 2025 paper (https://arxiv.org/abs/2505.20047). We shall soon open source that code!

Paper and discussion history. A mention is not an endorsement.

Grammars of Formal Uncertainty: When to Trust LLMs in Automated Reasoning Tasks — mentioned by barthelomew

Sat, 04 Oct 2025 19:46:17 GMT

Hey there! I mostly designed and wrote most of the actual interpreter during my internship at Microsoft Research last summer. Constrained decoding for GPT-4 wasn’t available when we started designing the DSL, and besides, creating a regex to constrain this specific DSL is quite challenging. When the grammar of the language is better defined, like SMT (https://arxiv.org/abs/2505.20047) - we are able to do this with open source LLMs.

Paper and discussion history. A mention is not an endorsement.

Grammars of Formal Uncertainty: When to Trust LLMs in Automated Reasoning Tasks — mentioned by barthelomew

Sat, 04 Oct 2025 19:39:22 GMT

The autoformalization gap is pretty difficult to bridge indeed. We explored uncertainty quantification of autoformalization on well-defined grammars in our NeurIPS 2025 paper : https://arxiv.org/abs/2505.20047 . If you ever feel like chatting and discussing more details, happy to chat!

Paper and discussion history. A mention is not an endorsement.

Grammars of Formal Uncertainty: When to Trust LLMs in Automated Reasoning Tasks — HN submission

Tue, 27 May 2025 02:27:19 GMT

Grammars of Formal Uncertainty

34 points; 5 discussion replies on HN. Paper and source history. A submission is not an endorsement.