1-23 of about 23 matches for site:arxiv.org seeking
https://arxiv.org/abs/2502.15989
2502.15989] Mean-Shift Distillation for Diffusion Mode Seeking Skip to main content We gratefully acknowledge support
https://arxiv.org/abs/2502.15989
2502.15989] Mean-Shift Distillation for Diffusion Mode Seeking Skip to main content We gratefully acknowledge support
https://arxiv.org/abs/2109.06157
in existing QA datasets. We find that a significant proportion of information seeking questions have context-dependent
https://arxiv.org/abs/2501.05445
strides in distilling image-generative models for 3D generation. However, its maximum-likelihood-seeking behavior often leads to
[2304.03279] Do the Rewards Justify the Means? Measuring Trade-Offs Between Rewards and Ethical Beha
https://arxiv.org/abs/2304.03279
Abstract: Artificial agents have traditionally been trained to maximize reward, which may incentivize power-seeking and deception
[2203.00130] Paper Plain: Making Medical Research Papers Approachable to Healthcare Consumers with N
https://arxiv.org/abs/2203.00130
Natural Language Processing, by Tal August and 4 other authors View PDF Abstract: When seeking information not covered in
https://arxiv.org/abs/2501.05445
strides in distilling image-generative models for 3D generation. However, its maximum-likelihood-seeking behavior often leads to
https://arxiv.org/abs/2405.14380
out a fit with data fully restricted to the dijet region seeking to minimize
[2304.03279] Do the Rewards Justify the Means? Measuring Trade-Offs Between Rewards and Ethical Beha
https://arxiv.org/abs/2304.03279
Abstract: Artificial agents have traditionally been trained to maximize reward, which may incentivize power-seeking and deception
https://arxiv.org/abs/2209.00626
goals which generalize beyond their fine-tuning distributions, and pursue those goals using power-seeking strategies. We review emerging
https://arxiv.org/abs/2310.01405
on a wide range of safety-relevant problems, including honesty, harmlessness, power-seeking, and more
https://arxiv.org/abs/2411.09222
substantively pluralistic, human-centered, participatory, and public-interest AI, (ii) can help guide organizations seeking to increase
https://arxiv.org/abs/2407.12883
and 14 other authors View PDF Abstract: Existing retrieval benchmarks primarily consist of information-seeking queries (e.g., aggregated questions
https://arxiv.org/abs/2508.17230
hard to extract a universal 3D representation from web datasets. Instead, we are seeking a general
https://arxiv.org/abs/2307.11049
careful design of reward functions or the use of novelty-seeking exploration bonuses. Human supervisors
https://arxiv.org/abs/2304.09848
for systems that may serve as a primary tool for information-seeking users, especially given their
https://arxiv.org/abs/2310.01405
on a wide range of safety-relevant problems, including honesty, harmlessness, power-seeking, and more
https://arxiv.org/abs/2307.11049
careful design of reward functions or the use of novelty-seeking exploration bonuses. Human supervisors
https://arxiv.org/abs/2202.14020
distribution, and can only be applied to images generated by StyleGAN itself. Seeking to bring
https://arxiv.org/abs/2202.14020
distribution, and can only be applied to images generated by StyleGAN itself. Seeking to bring
https://arxiv.org/abs/2406.12137
chat session with Claude 3), and associated information is accessible to parties seeking to interact