jaspreet paper

_bibliography/papers.bib

@@ -1,5 +1,16 @@
 ---
 ---
+@misc{ranjit2024oath,
+      title={OATH-Frames: Characterizing Online Attitudes Towards Homelessness via LLM Assistants},
+      author={Jaspreet Ranjit and Brihi Joshi and Rebecca Dorn and Laura Petry and Olga Koumoundouros and Jayne Bottarini and Peichen Liu and Eric Rice and Swabha Swayamdipta},
+      year={2024},
+      abbr={Preprint},
+      url={../assets/pdf/papers/oath_frames.pdf},
+      selected=true,
+      preview={oath_frames_dist.png},
+      selected=true,
+      abstract={Homelessness in the U.S. is widespread; individual beliefs and attitudes towards homelessness—often expressed on social media are complex and nuanced (e.g. critical as well as sympathetic). Such attitudes can be challenging to summarize at scale, obfuscating the broader public opinion which advocacy organizations use to guide public policy and reform efforts. Our work proposes an approach to enable a large-scale study on homelessness via two major contributions. First, with the help of domain experts in social work and their trainees, we characterize Online Attitudes towards Homelessness in nine hierarchical frames (OATH-Frames) on a collection of 4K social media posts. Further, in an effort to ease the annotation of these frames, we employ GPT-4 as an LLM assistant to the experts; GPT-4 + Expert annotation presents an attractive trade off owing to a 6.5× speedup in annotation time despite only incurring a 2 point F1 difference in annotation performance. Our effort results in a collection of 8K social media posts labeled by domain and trained experts (with and without GPT-4 assistance). Second, using predicted OATH-Frames on a Flan-T5-Large model trained on our data, we perform a large-scale analysis on 2.4M posts on homelessness. We find that posts that contain mentions of west coast states express more harmful generalizations of people experiencing homelessness (PEH) compared to posts about east coast states. We also find marked differences in attitudes across vulnerable populations as they are compared to PEH as being either more or less deserving of aid.},
+}
 
 @misc{finlayson2024logits,
       title={Logits of API-Protected LLMs Leak Proprietary Information},
@@ -12,15 +23,6 @@ @misc{finlayson2024logits
       abstract={The commercialization of large language models (LLMs) has led to the common practice of high-level API-only access to proprietary models. In this work, we show that even with a conservative assumption about the model architecture, it is possible to learn a surprisingly large amount of non-public information about an API-protected LLM from a relatively small number of API queries (e.g., costing under $1,000 for OpenAI's gpt-3.5-turbo). Our findings are centered on one key observation: most modern LLMs suffer from a softmax bottleneck, which restricts the model outputs to a linear subspace of the full output space. We show that this lends itself to a model image or a model signature which unlocks several capabilities with affordable cost: efficiently discovering the LLM's hidden size, obtaining full-vocabulary outputs, detecting and disambiguating different model updates, identifying the source LLM given a single full LLM output, and even estimating the output layer parameters. Our empirical investigations show the effectiveness of our methods, which allow us to estimate the embedding size of OpenAI's gpt-3.5-turbo to be about 4,096. Lastly, we discuss ways that LLM providers can guard against these attacks, as well as how these capabilities can be viewed as a feature (rather than a bug) by allowing for greater transparency and accountability.},
 }
 
-@misc{cui2024annotating,
-      title={Annotating FrameNet via Structure-Conditioned Language Generation},
-      author={Xinyue Cui and Swabha Swayamdipta},
-      pdf={papers/fn_conditioned_generation.pdf},
-      abbr={Preprint},
-      year={2024},
-      abstract={Despite the mounting evidence for generative  capabilities of language models in understanding and generating natural language, their effectiveness on explicit manipulation and generation of linguistic structures remain understudied. In this paper, we investigate the task of generating new sentences preserving a given semantic structure, following the FrameNet formalism. We propose a framework to produce novel frame-semantically annotated sentences following an overgenerate-and-filter approach. Our results show that conditioning on rich, explicit semantic information tends to produce generations with high human acceptance, under both prompting and finetuning. Nevertheless, we discover that generated frame-semantic structured data is ineffective at training data augmentation for frame-semantic role labeling. Our study concludes that while generating high-quality, semantically rich data might be within reach, their downstream utility remains to be seen, highlighting the outstanding challenges with automating linguistic annotation tasks.},
-}
-
 @misc{nazari2024generative,
       title={Generative Explanations for Program Synthesizers},
       author={Amirmohammad Nazari and Souti Chattopadhyay and Swabha Swayamdipta and Mukund Raghothaman},

_bibliography/papers_orcid.bib

@@ -11,7 +11,11 @@ @misc{finlayson2024logits
 @misc{cui2024annotating,
       title={Annotating FrameNet via Structure-Conditioned Language Generation},
       author={Xinyue Cui and Swabha Swayamdipta},
+      url={../assets/pdf/papers/fn_conditioned_generation.pdf},
+      abbr={Preprint},
       year={2024},
+      preview={fn-conditioned-generation.jpeg},
+      abstract={Despite the mounting evidence for generative  capabilities of language models in understanding and generating natural language, their effectiveness on explicit manipulation and generation of linguistic structures remain understudied. In this paper, we investigate the task of generating new sentences preserving a given semantic structure, following the FrameNet formalism. We propose a framework to produce novel frame-semantically annotated sentences following an overgenerate-and-filter approach. Our results show that conditioning on rich, explicit semantic information tends to produce generations with high human acceptance, under both prompting and finetuning. Nevertheless, we discover that generated frame-semantic structured data is ineffective at training data augmentation for frame-semantic role labeling. Our study concludes that while generating high-quality, semantically rich data might be within reach, their downstream utility remains to be seen, highlighting the outstanding challenges with automating linguistic annotation tasks.},
 }
 
 @misc{nazari2024generative,

_pages/about.md

@@ -23,8 +23,8 @@ My goal is to design frameworks that allow robust, and reliable frameworks that
 - Understanding the Mechanisms that Drive Language Technologies:
   : Even the most reliable evaluation may not reveal much about the mechanisms driving powerful yet opaque models. What do model geometries reveal about the processes underlying our models, and how can we improve models through different designs? Are models by design limited to making some choices which can uniquely identify them?
 
-- Human-and-AI Collaboration:
-  : AI technologies are designed by humans and for humans and the future of these technologies involves cooperation with humans. How can we reliably say when a model serves the general-purpose or custom utility for a human user? Where can these technologies complement human capabilities and where not?
+- Human and AI Collaboration:
+  : AI technologies are designed by humans and for humans, the future of AI involves cooperation and collaboration with humans. How can we reliably say when a general-purpose model will serve the custom utility for a human user? Where can these technologies complement human capabilities and where not?
 
 These challenges require novel and creative techniques for redesigning generative evaluation to keep pace with model performance. This brings together a broad array of empirical research with theoretical fundamentals underlying language models.