Deductive Closure Training of Language Models for Coherence, Accuracy, and Updatability
fine-tuning a language model on the implications of its own text that survive a self-consistency check
Afra Feyza Akyürek, Ekin Akyürek, Leshem Choshen, Derry Wijaya, Jacob Andreas · Findings of ACL 2024 · 2024
In one sentence
Deductive Closure Training (DCT) prompts a language model to generate the implications and contradictions of seed text, keeps the most probable logically consistent subset, and fine-tunes on it — turning inference-time reasoning into training-time supervision for factuality and model updating.
Abstract
While language models (LMs) can sometimes generate factually correct text and estimate truth values of individual claims, these generally do not reflect a globally coherent, manipulable model of the world. As a consequence, current LMs also generate incorrect or nonsensical content, and are difficult to edit and bring up to date. We present a method called Deductive Closure Training (DCT) that uses LMs themselves to identify implications of (and contradictions within) the text that they generate, yielding an efficient self-supervised procedure for improving LM factuality. Given a collection of seed documents, DCT prompts LMs to generate additional text implied by these documents, reason globally about the correctness of this generated text, and finally fine-tune on text inferred to be correct. Given seed documents from a trusted source, DCT provides a tool for supervised model updating; if seed documents are sampled from the LM itself, DCT enables fully unsupervised fine-tuning for improved coherence and accuracy. Across the CREAK, MQUaKE, and Reversal Curse datasets, supervised DCT improves LM fact verification and text generation accuracy by 3-26%; on CREAK fully unsupervised DCT improves verification accuracy by 12%. These results show that LMs' reasoning capabilities during inference can be leveraged during training to improve their reliability.
Questions this paper answers
- can a language model get more accurate about facts by training on statements it wrote itself, with no human labels?
- how much does unsupervised fine-tuning on self-generated, consistency-filtered statements improve fact-verification accuracy on CREAK?
- how do I raise a 7B model's factual accuracy when I have no labelled fact-checking data?
- is self-generated training data enough to beat few-shot prompting for fact verification on my own model?
- Unsupervised Deductive Closure Training raises Llama-2-7B fact-verification accuracy on the CREAK validation set from 71.7% with few-shot prompting to 83.5%, using no labelled data at all.
Holds for: Llama-2-7B on CREAK validation, averaged over 3 seeds; seeds sampled from the model itself at temperature 0.9, LoRA fine-tuning for 30 epochs, no held-out dev set.
- Dropping the logical consistency-evaluation step from unsupervised DCT lowers CREAK accuracy from 83.5% to 77.5%, below the 80.0% obtained by fine-tuning on seed statements alone.
Holds for: Llama-2-7B on CREAK validation, 3 seeds, same 586 generated documents in both the full and ablated conditions.
- if a model trains on text it generated itself, does anything filter out the false parts?
- what is the contribution of the logical consistency-evaluation step in Deductive Closure Training, relative to fine-tuning on seed statements alone?
- do I need to check generated implications for consistency before fine-tuning on them, or can I train on all of them?
- can I skip the consistency-checking stage and still come out ahead of just fine-tuning on the seed statements?
- Dropping the logical consistency-evaluation step from unsupervised DCT lowers CREAK accuracy from 83.5% to 77.5%, below the 80.0% obtained by fine-tuning on seed statements alone.
Holds for: Llama-2-7B on CREAK validation, 3 seeds, same 586 generated documents in both the full and ablated conditions.
- A double-checking pass in which the model re-verifies each generated implication or contradiction raises their validity rate from 76% to 82%. It also discards about 2/3 of generations, cutting 100 annotated statements to 33.
Holds for: Manual annotation of 100 implications and contradictions generated by Llama-2-7B in the supervised CREAK setting; precision is bought at a large loss of training data volume.
- if I have only a couple dozen labelled examples, is it worth combining them with machine-generated training data?
- does pooling a small labelled set with unsupervised Deductive Closure Training data beat both supervised fine-tuning and the unsupervised setting on CREAK?
- how do I make 20 hand-labelled claims go further for training a fact-verification model?
- I have about 20 annotated claims and lots of unlabelled text — should I train on both together?
- Pooling 20 labelled CREAK claims with unsupervised DCT data reaches 84.9% accuracy, above both supervised fine-tuning on the 20 claims (77.2%) and unsupervised DCT alone (83.5%).
Holds for: Llama-2-7B, CREAK validation, 3 seeds; supervised seeds are 20 training claims filtered to those labelled true, with supervised runs trained for 60 epochs.
- does letting a model see the unlabelled test questions during training help it answer them correctly?
- how does transductive Deductive Closure Training on unlabelled CREAK validation claims compare with inference-time implication-graph reasoning?
- I have the unlabelled evaluation claims in hand — should I fine-tune on their implications or run consistency reasoning at inference time?
- is it better to bake consistency reasoning into the weights or pay for it at inference on every query?
- With unlabelled CREAK validation claims used as seed documents, DCT on related claims plus double-checking reaches 85.5% accuracy. The inference-time Graph-Inference baseline, which reasons over implication graphs without fine-tuning, reaches 77.7%.
Holds for: Llama-2-7B, all 1,371 unlabelled CREAK validation claims used as seeds, 1 training epoch; requires access to test-set inputs, so it is not a strict train/test split.
- can training a model on its own statements stop it from agreeing with two claims that contradict each other?
- does self-training on generated contradictions reduce the rate of contradictory statement pairs both labelled true, and what happens to accuracy?
- how do I stop a model from labelling a claim and its negation as both true?
- if I fine-tune for logical coherence, do I lose accuracy to get consistency?
- On the 250-pair CREAK contrast set, unsupervised DCT cuts the fraction of contradictory statement pairs labelled both-true from 34.4% to 19.2% while raising accuracy from 63.2% to 73.0%; the contradiction-only variant drives both-true down furthest, to 12.0%.
Holds for: Llama-2-7B, CREAK contrast set of 250 lexically similar pairs with opposite truth values; coherence measured only as the both-true rate, not as full logical consistency.
- after you correct a fact in a language model, why does it still get wrong the questions that depend on that fact?
- how does Deductive Closure Training compare with MEMIT, MeLLo and fine-tuning on edits for multi-hop counterfactual accuracy on MQuAKE?
- how do I edit thousands of facts in a model so that multi-hop answers depending on them also change?
- which knowledge-editing approach should I pick if my questions require chaining two edited facts together?
- On the MQuAKE counterfactual multi-hop subset, DCT with combined correlative and standard implications reaches 35.6% accuracy at 20 edits and 15.4% at 1,000 edits. The corresponding numbers are 18.3% and 11.4% for MeLLo, 11.7% and 0.6% for MEMIT, and 9.0% and 4.1% for fine-tuning on edits.
Holds for: Llama-2-7B, MQuAKE-CF multi-hop questions only, 3 seeds except the 1,000-edit setting which is run once; each edit yields 3 test questions and edits are converted to questions by prompting before fine-tuning.
- Prompting the model to state related background facts before generating implications (correlative implications) beats plain implications on MQuAKE at every edit count, for example 41.7% versus 20.0% at 10 edits.
Holds for: Llama-2-7B on MQuAKE-CF multi-hop questions, edit counts 10–1,000; the advantage of correlative over plain implications is systematic but the combined set is best on average.
- does asking a model to recall related background facts first make the statements it generates more useful for training?
- do correlative implications, generated after eliciting related background knowledge, outperform plain implications on MQuAKE and introduce more new information?
- how should I prompt a model to generate training statements that carry consequences of an edited fact rather than restating it?
- is it worth an extra prompting step to get background facts before generating implications of an edit?
- Prompting the model to state related background facts before generating implications (correlative implications) beats plain implications on MQuAKE at every edit count, for example 41.7% versus 20.0% at 10 edits.
Holds for: Llama-2-7B on MQuAKE-CF multi-hop questions, edit counts 10–1,000; the advantage of correlative over plain implications is systematic but the combined set is best on average.
- Correlative implications on MQuAKE incorporate new information supplied by the model 33% of the time versus 24% for plain implications, about 50% more. Only 30–36% of generations in either condition are strict logical implications.
Holds for: Manual annotation of roughly 350 generations from Llama-2-7B across experiments; single-annotator qualitative analysis, not a validated protocol.
- if a model learns that A is B, can training fix the fact that it cannot answer the question the other way round?
- does fine-tuning on generated implications mitigate the reversal curse, and at what cost to forward-direction exact-match accuracy?
- how do I teach a model a relation so it answers questions in both directions rather than only the one it was trained on?
- will training on generated implications fix reversed-order recall without degrading the direction that already worked?
- On the Reversal Curse child-to-parent split, DCT raises exact-match accuracy on reversed questions from 2.2% to 48.3% while keeping same-direction accuracy at 87.9% versus 95.3% for ordinary fine-tuning.
Holds for: Llama-2-7B, 1,000 child-to-parent evaluation examples; gains are far smaller on person-to-description (3.7% to 10.7%) and description-to-person (3.0% to 15.7%), 300 examples each.
- is there any proof that training a model on text it produced itself should make it more accurate rather than reinforce its mistakes?
- under what assumptions on seed-pair correctness and conditional probability gain is Deductive Closure Training guaranteed to increase the probability of the correct answer?
- is there a formal argument I can rely on before I fine-tune a model on its own generations?
- Deductive Closure Training is proved to strictly increase the probability a language model assigns to the correct answer for a question. The proof assumes self-generated seed question-answer pairs are correct with probability at least p*, and that conditioning on a correct related pair raises correct-answer probability by at least 1/p*.
Holds for: A minimal unsupervised variant of DCT with a single implication per seed, no consistency check, and training to convergence on arbitrarily many seeds; the two assumptions are not empirically verified.
- if a model re-checks the statements it generated, do they get more reliable, and how much gets thrown away?
- how much does a double-checking verification pass raise the validity rate of generated implications and contradictions, and what fraction is discarded?
- how do I filter self-generated implications for validity without discarding most of my training set?
- is a second verification pass over generated statements worth losing most of the data I generated?
- A double-checking pass in which the model re-verifies each generated implication or contradiction raises their validity rate from 76% to 82%. It also discards about 2/3 of generations, cutting 100 annotated statements to 33.
Holds for: Manual annotation of 100 implications and contradictions generated by Llama-2-7B in the supervised CREAK setting; precision is bought at a large loss of training data volume.
- which work first turned checking a claim against its logical consequences into a way of training a model rather than a way of answering at inference time?
- which paper reframes inference-time consistency reasoning over implications and contradictions as training-time supervision for factuality?
- where should I start reading about improving a language model's factuality without labelled data?
- Deductive Closure Training reframes inference-time consistency reasoning over implications and contradictions of a claim as a source of training-time supervision. It gives one framework for both unsupervised factuality improvement and supervised model editing.
Holds for: Positioned as of publication at ACL Findings 2024 relative to inference-time belief-graph methods and to concurrent self-training work that ranks or filters self-generated data by model-assigned factuality labels rather than extrapolating implications.
- Deductive Closure Training is an entry point for readers asking whether model editing propagates to a fact's consequences rather than to the edited sentence alone. It is evaluated on multi-hop question answering and reversal-style generalization.
Holds for: Demonstrated only on English-language datasets (CREAK, MQuAKE-CF, Reversal Curse) with a single Llama-2-7B model, and only with one layer of implications rather than a deeper graph of relations.
- what should I read about whether correcting a fact in a language model also corrects everything that follows from it?
- which work evaluates whether knowledge edits propagate to multi-hop questions and to reversed relation directions?
- how do I find out whether an edit to one fact in a model generalizes to its consequences before I rely on it?
- Deductive Closure Training is an entry point for readers asking whether model editing propagates to a fact's consequences rather than to the edited sentence alone. It is evaluated on multi-hop question answering and reversal-style generalization.
Holds for: Demonstrated only on English-language datasets (CREAK, MQuAKE-CF, Reversal Curse) with a single Llama-2-7B model, and only with one layer of implications rather than a deeper graph of relations.
- On the MQuAKE counterfactual multi-hop subset, DCT with combined correlative and standard implications reaches 35.6% accuracy at 20 edits and 15.4% at 1,000 edits. The corresponding numbers are 18.3% and 11.4% for MeLLo, 11.7% and 0.6% for MEMIT, and 9.0% and 4.1% for fine-tuning on edits.
Holds for: Llama-2-7B, MQuAKE-CF multi-hop questions only, 3 seeds except the 1,000-edit setting which is run once; each edit yields 3 test questions and edits are converted to questions by prompting before fine-tuning.
Claims and scope
- Unsupervised Deductive Closure Training raises Llama-2-7B fact-verification accuracy on the CREAK validation set from 71.7% with few-shot prompting to 83.5%, using no labelled data at all. (Table 1)
Scope: Llama-2-7B on CREAK validation, averaged over 3 seeds; seeds sampled from the model itself at temperature 0.9, LoRA fine-tuning for 30 epochs, no held-out dev set.
- Dropping the logical consistency-evaluation step from unsupervised DCT lowers CREAK accuracy from 83.5% to 77.5%, below the 80.0% obtained by fine-tuning on seed statements alone. (Table 1)
Scope: Llama-2-7B on CREAK validation, 3 seeds, same 586 generated documents in both the full and ablated conditions.
- Pooling 20 labelled CREAK claims with unsupervised DCT data reaches 84.9% accuracy, above both supervised fine-tuning on the 20 claims (77.2%) and unsupervised DCT alone (83.5%). (Table 1)
Scope: Llama-2-7B, CREAK validation, 3 seeds; supervised seeds are 20 training claims filtered to those labelled true, with supervised runs trained for 60 epochs.
- With unlabelled CREAK validation claims used as seed documents, DCT on related claims plus double-checking reaches 85.5% accuracy. The inference-time Graph-Inference baseline, which reasons over implication graphs without fine-tuning, reaches 77.7%. (Table 1)
Scope: Llama-2-7B, all 1,371 unlabelled CREAK validation claims used as seeds, 1 training epoch; requires access to test-set inputs, so it is not a strict train/test split.
- On the 250-pair CREAK contrast set, unsupervised DCT cuts the fraction of contradictory statement pairs labelled both-true from 34.4% to 19.2% while raising accuracy from 63.2% to 73.0%; the contradiction-only variant drives both-true down furthest, to 12.0%. (Table 2)
Scope: Llama-2-7B, CREAK contrast set of 250 lexically similar pairs with opposite truth values; coherence measured only as the both-true rate, not as full logical consistency.
- On the MQuAKE counterfactual multi-hop subset, DCT with combined correlative and standard implications reaches 35.6% accuracy at 20 edits and 15.4% at 1,000 edits. The corresponding numbers are 18.3% and 11.4% for MeLLo, 11.7% and 0.6% for MEMIT, and 9.0% and 4.1% for fine-tuning on edits. (Table 3)
Scope: Llama-2-7B, MQuAKE-CF multi-hop questions only, 3 seeds except the 1,000-edit setting which is run once; each edit yields 3 test questions and edits are converted to questions by prompting before fine-tuning.
- Prompting the model to state related background facts before generating implications (correlative implications) beats plain implications on MQuAKE at every edit count, for example 41.7% versus 20.0% at 10 edits. (Table 3)
Scope: Llama-2-7B on MQuAKE-CF multi-hop questions, edit counts 10–1,000; the advantage of correlative over plain implications is systematic but the combined set is best on average.
- Correlative implications on MQuAKE incorporate new information supplied by the model 33% of the time versus 24% for plain implications, about 50% more. Only 30–36% of generations in either condition are strict logical implications. (Appendix C)
Scope: Manual annotation of roughly 350 generations from Llama-2-7B across experiments; single-annotator qualitative analysis, not a validated protocol.
- On the Reversal Curse child-to-parent split, DCT raises exact-match accuracy on reversed questions from 2.2% to 48.3% while keeping same-direction accuracy at 87.9% versus 95.3% for ordinary fine-tuning. (Table 4)
Scope: Llama-2-7B, 1,000 child-to-parent evaluation examples; gains are far smaller on person-to-description (3.7% to 10.7%) and description-to-person (3.0% to 15.7%), 300 examples each.
- A double-checking pass in which the model re-verifies each generated implication or contradiction raises their validity rate from 76% to 82%. It also discards about 2/3 of generations, cutting 100 annotated statements to 33. (Appendix C)
Scope: Manual annotation of 100 implications and contradictions generated by Llama-2-7B in the supervised CREAK setting; precision is bought at a large loss of training data volume.
- Deductive Closure Training is proved to strictly increase the probability a language model assigns to the correct answer for a question. The proof assumes self-generated seed question-answer pairs are correct with probability at least p*, and that conditioning on a correct related pair raises correct-answer probability by at least 1/p*. (Section 4 and Appendix E)
Scope: A minimal unsupervised variant of DCT with a single implication per seed, no consistency check, and training to convergence on arbitrarily many seeds; the two assumptions are not empirically verified.
- Deductive Closure Training reframes inference-time consistency reasoning over implications and contradictions of a claim as a source of training-time supervision. It gives one framework for both unsupervised factuality improvement and supervised model editing. (Section 2)
Scope: Positioned as of publication at ACL Findings 2024 relative to inference-time belief-graph methods and to concurrent self-training work that ranks or filters self-generated data by model-assigned factuality labels rather than extrapolating implications.
- Deductive Closure Training is an entry point for readers asking whether model editing propagates to a fact's consequences rather than to the edited sentence alone. It is evaluated on multi-hop question answering and reversal-style generalization. (Section 5)
Scope: Demonstrated only on English-language datasets (CREAK, MQuAKE-CF, Reversal Curse) with a single Llama-2-7B model, and only with one layer of implications rather than a deeper graph of relations.
Common misreadings
- Deductive Closure Training does not solve the Reversal Curse: with the same general-purpose implication prompt, reversed-direction accuracy on the person-to-description split rises only from 3.7% to 10.7%, and the paper describes the challenge as far from solved.
- The accuracy gains are not simply the effect of fine-tuning on more self-generated text: removing the consistency-evaluation step drops CREAK accuracy from 83.5% to 77.5% with the identical 586 generated documents.
- Most generated statements are not strict deductions; manual annotation found only 30–36% of MQuAKE generations to be strict logical implications, with the rest plausible-but-not-entailed text.
- The transductive CREAK result of 85.5% is not comparable to a standard train/test result, because it uses the unlabelled validation claims themselves as seed documents.
- The formal guarantee is not a general proof that self-training improves language models: it holds for a minimal single-implication variant without the consistency check, and only when two stated assumptions about seed correctness and prompt benefit are satisfied.
- All results come from a single model, Llama-2-7B, and English-only datasets, so the reported gains are not evidence about larger models or other languages.
Terminology in this paper
- Deductive closure
- The complete set of statements that follow from a given collection of statements; a language model is deductively closed if it assigns high probability to a logically coherent set of assertions and to all their implications.
- Correlative implications
- Implications generated after first prompting the model to list related background facts about the entities in a seed claim, so that the resulting statements combine the new information with the model's existing knowledge.
- Consistency evaluation
- Selecting, among all assignments of True/False to a seed document and its generated implications and contradictions, the most probable assignment that respects every implication and contradiction.
- Transductive DCT
- Applying Deductive Closure Training with unlabelled inputs from the evaluation set used as seed documents, so the model is updated with access to the test claims but not their labels.
- Double-checking
- A filtering pass in which the language model is re-prompted to verify that a seed statement really does entail or contradict each generated statement, discarding those it does not confirm.
How to cite
@inproceedings{DBLP:conf/acl/AkyurekACWA24,author = {Afra Feyza Aky{\"{u}}rek and
Ekin Aky{\"{u}}rek and
Leshem Choshen and
Derry Wijaya and
Jacob Andreas},
editor = {Lun{-}Wei Ku and
Andre Martins and
Vivek Srikumar},
title = {Deductive Closure Training of Language Models for Coherence, Accuracy,
and Updatability},
booktitle = {Findings of the Association for Computational Linguistics, {ACL} 2024,
Bangkok, Thailand and virtual meeting, August 11-16, 2024},
pages = {9802--9818},
publisher = {Association for Computational Linguistics},
year = {2024},
url = {https://aclanthology.org/2024.findings-acl.584},
timestamp = {Tue, 27 Aug 2024 17:38:11 +0200},
biburl = {https://dblp.org/rec/conf/acl/AkyurekACWA24.bib},
bibsource = {dblp computer science bibliography, https://dblp.org}
}
References
See the full reference list in the paper.