--- title: "Dan Jurafsky — papers, page 2 · Stanford University" description: "Dan Jurafsky, Stanford University: 49 papers on Elamur Research (1998–2026), h-index 85 — affiliations, topics, co-authors and who cites…" canonical: "https://research.elamur.ai/free/en/authors/dan-jurafsky/page/2/" --- [Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/authors/dan-jurafsky/page/2/) [home](https://md.research.elamur.ai/free/en/) / [authors](https://md.research.elamur.ai/free/en/authors/) / [Dan Jurafsky](https://md.research.elamur.ai/free/en/authors/dan-jurafsky/) / page 2 # Dan Jurafsky · page 2 49 papers in our library · 1998–2026 · page 2 of 2 [The profile, affiliations and co-authors →](https://md.research.elamur.ai/free/en/authors/dan-jurafsky/) ## Papers The most cited first (OpenAlex and our library). [← prev](https://md.research.elamur.ai/free/en/authors/dan-jurafsky/) [1](https://md.research.elamur.ai/free/en/authors/dan-jurafsky/) **2** 26. **[Marked Personas: Using Natural Language Prompts to Measure Stereotypes in Language Models](https://md.research.elamur.ai/free/en/papers/cheng-2023-marked-personas-natural-language-prompts/)** *(2023 · 246 citations)* The paper presents Marked Personas, a way to measure stereotypes in large language models without building a lexicon or labeling data. 27. **[Framing and Agenda-setting in Russian News: a Computational Analysis of Intricate Political Strategies](https://md.research.elamur.ai/free/en/papers/field-2018-framing-agenda-setting-russian-news/)** *(2018 · 195 citations)* Most NLP work on state media manipulation targets fabricated stories and censorship. This paper looks at quieter tactics, using two political-science concepts: agenda-setting (which topics get covered) and framing (which aspects of a topic are emphasized). The authors study more than 100K articles from the Russian… 28. **[ELEPHANT: Measuring and understanding social sycophancy in LLMs](https://md.research.elamur.ai/free/en/papers/cheng-2025-elephant-measuring-understanding-social-sycophancy/)** *(2025 · 188 citations)* Earlier sycophancy research checks whether an LLM drops a factual answer to agree with a belief the user states outright. 29. **[Knowledge-Free Induction of Inflectional Morphologies](https://md.research.elamur.ai/free/en/papers/schone-2001-knowledge-free-induction-inflectional-morphologies/)** *(2001 · 170 citations)* The paper induces the morphology of an inflectional language from raw text, with no human input. The algorithm combines three cues: orthography, semantics, and syntactic distributions. Its output is a set of conflation sets, groups of the inflected and derived forms of one word. The method is applied to German, Dutch… 30. **[Neural Language Correction with Character-Based Attention](https://md.research.elamur.ai/free/en/papers/xie-2016-neural-language-correction-character-based/)** *(2016 · 164 citations)* The paper addresses automatic correction of English learners' writing. Earlier systems either relied on per-error-type classifiers, which are precise but miss redundancy and unidiomatic wording, or on phrase-based machine translation, which handles misspellings poorly. The authors cast correction as character-level… 31. **[Parsing to Stanford Dependencies: Trade-offs between Speed and Accuracy](https://md.research.elamur.ai/free/en/papers/cer-2010-parsing-stanford-dependencies-trade-offs/)** *(2010 · 158 citations)* The authors compare ways of producing Stanford Dependencies quickly. One group is constituency parsers (Stanford, Charniak, Charniak-Johnson reranker, Bikel, Berkeley) followed by a conversion to dependencies; the other is dependency parsers trained directly on the basic dependencies (MaltParser's Nivre variants… 32. **[Neural Net Models of Open-domain Discourse Coherence](https://md.research.elamur.ai/free/en/papers/li-2017-neural-net-models-open-domain/)** *(2017 · 147 citations)* Discourse coherence, meaning how sentences connect into a sensible whole, matters for both understanding and generating text. 33. **[A computational approach to politeness with application to social factors](https://md.research.elamur.ai/free/en/papers/mizil-2013-computational-approach-politeness-application-social/)** *(2013 · 141 citations)* The paper addresses how to detect politeness in text automatically. It focuses on requests, because asking something of another person is an imposition and tends to trigger politeness strategies. The authors collected requests from Wikipedia user talk-pages and Stack Exchange comments. They had a subset rated on… 34. **[Safety-Tuned LLaMAs: Lessons From Improving the Safety of Large Language Models that Follow Instructions](https://md.research.elamur.ai/free/en/papers/bianchi-2023-safety-tuned-llamas-lessons-improving/)** *(2023 · 135 citations)* Open instruction-tuned LLMs are usually optimized purely for helpfulness, so they also comply with harmful requests. 35. **[Pile of Law: Learning Responsible Data Filtering from the Law and a 256GB Open-Source Legal Dataset](https://md.research.elamur.ai/free/en/papers/henderson-2022-pile-law-learning-responsible-data/)** *(2022 · 133 citations)* Teams that pretrain large language models remove private or toxic text in inconsistent, improvised ways and usually ignore context. 36. **[Lexical, Prosodic, and Syntactic Cues for Dialog Acts](https://md.research.elamur.ai/free/en/papers/jurafsky-1998-lexical-prosodic-syntactic-cues-dialog/)** *(1998 · 132 citations)* The paper examines how short feedback utterances in conversation, such as yeah , uh-huh and okay , can be told apart by their dialog act (DA) function. 37. **[ReFT: Representation Finetuning for Language Models](https://md.research.elamur.ai/free/en/papers/wu-2024-reft-representation-finetuning-language-models/)** *(2024 · 74 citations)* Most parameter-efficient finetuning (PEFT) techniques, such as LoRA, adapt a model by changing a small set of its weights. 38. **[Dialect prejudice predicts AI decisions about people's character, employability, and criminality](https://md.research.elamur.ai/free/en/papers/hofmann-2024-dialect-prejudice-predicts-ai-decisions/)** *(2024 · 57 citations)* Earlier bias research on language models mostly asked about racial groups by name. This paper asks whether models also hold hidden racial prejudice that is triggered only by how a person writes, specifically African American English (AAE) compared with Standardized American English (SAE). The authors adapt the… 39. **[Emergence of Gricean Maxims from Multi-Agent Decision Theory](https://md.research.elamur.ai/free/en/papers/vogel-2013-emergence-gricean-maxims-multi-agent/)** *(2013 · 51 citations)* Grice's cooperative principle and maxims (quality, relation, quantity, manner) are usually stated as norms. 40. **[Implicatures and Nested Beliefs in Approximate Decentralized-POMDPs](https://md.research.elamur.ai/free/en/papers/vogel-2013-implicatures-nested-beliefs-approximate-decentralized/)** *(2013 · 45 citations)* Gricean implicatures require agents to reason about what others believe about their own beliefs. Such nested reasoning is hard to model computationally. Vogel, Potts and Jurafsky treat communication as a decision problem inside a Decentralized-POMDP (Dec-POMDP), in which two agents share a reward. Exact Dec-POMDP… 41. **[CausalGym: Benchmarking causal interpretability methods on linguistic tasks](https://md.research.elamur.ai/free/en/papers/arora-2024-causalgym-benchmarking-causal-interpretability-methods/)** *(2024 · 38 citations)* CausalGym connects two research lines that have mostly stayed apart. Psycholinguistic work tests language models on large behavioural benchmarks. Interpretability work tries to find internal causal mechanisms but usually evaluates on small, one-off datasets. The authors turn SyntaxGym test suites into templates that… 42. **[Dynamic Cheatsheet: Test-Time Learning with Adaptive Memory](https://md.research.elamur.ai/free/en/papers/suzgun-2026-dynamic-cheatsheet-test-time-learning/)** *(2026 · 37 citations)* Large language models usually treat each query as a fresh problem. They do not carry lessons from earlier attempts, so they rediscover solutions and repeat errors. The authors propose Dynamic Cheatsheet (DC), a method that gives a black-box LM an external memory that changes during inference. No weights are updated… 43. **[Focus on what matters: Applying Discourse Coherence Theory to Cross Document Coreference](https://md.research.elamur.ai/free/en/papers/held-2021-focus-what-matters-applying-discourse/)** *(2021 · 34 citations)* Cross-document coreference resolution (CDCR) for events and entities faces a quadratic explosion of candidate mention pairs. 44. **[Sensitivity as a Complexity Measure for Sequence Classification Tasks](https://md.research.elamur.ai/free/en/papers/hahn-2021-sensitivity-as-complexity-measure-sequence/)** *(2021 · 26 citations)* Why can bag-of-words models handle some NLP classification tasks, while others yielded only to large pretrained networks? Hahn, Jurafsky and Futrell answer this by adapting sensitivity, a concept from the analysis of Boolean functions, to natural-language inputs. 45. **[Detecting Institutional Dialog Acts in Police Traffic Stops](https://md.research.elamur.ai/free/en/papers/prabhakaran-2018-detecting-institutional-dialog-acts-police/)** *(2018 · 20 citations)* Police body-worn cameras capture large amounts of traffic-stop conversation, but most of it is never analysed. 46. **[Can Unconfident LLM Annotations Be Used for Confident Conclusions?](https://md.research.elamur.ai/free/en/papers/gligoric-2025-can-unconfident-llm-annotations-be/)** *(2025 · 19 citations)* Computational social scientists increasingly add LLM annotations to costly human labels. However, LLM errors and biases can quietly distort downstream statistics such as prevalences or regression coefficients. The authors propose CONFIDENCE-DRIVEN INFERENCE, which asks an LLM for both a label and a verbalized… 47. **[From Insanely Jealous to Insanely Delicious: Computational Models for the Semantic Bleaching of English Intensifiers](https://md.research.elamur.ai/free/en/papers/luo-2019-insanely-jealous-insanely-delicious-computational/)** *(2019 · 15 citations)* The paper presents what the authors call the first computational treatment of semantic bleaching, the process by which a word loses parts of its original meaning and becomes more abstract. 48. **[Attention to Non-Adopters](https://md.research.elamur.ai/free/en/papers/zhou-2026-attention-non-adopters/)** *(2026 · 1 citation)* This position paper argues that chat-model research mostly learns from people who already use LLMs, through interaction logs, preference data and user studies. 49. **[🧑‍🍳 Cooking Up Creativity: Enhancing LLM Creativity through Structured Recombination](https://md.research.elamur.ai/free/en/papers/mizrahi-2026-cooking-up-creativity-enhancing-llm/)** *(2026)* LLMs write fluent text but tend to repeat familiar patterns when asked for original ideas. Raising temperature only adds token-level randomness and does not produce new concepts. The authors move the creative step outside the model. An LLM turns textual ideas into structured representations, an algorithm recombines… [← prev](https://md.research.elamur.ai/free/en/authors/dan-jurafsky/) [1](https://md.research.elamur.ai/free/en/authors/dan-jurafsky/) **2**