Pith. sign in

REVIEW 12 cited by

FUDGE: Controlled Text Generation With Future Discriminators

Not yet reviewed by Pith; the record is open.

Re-run · record.json Download PDF Read on arXiv ↗

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2104.05218 v2 pith:OYOZHXSV submitted 2021-04-12 cs.CL cs.LG

Kevin Yang , Dan Klein This is my paper · ORCID
classification cs.CLcs.LG
keywords fudgegenerationattributetextcontrolleddesireddiscriminatorsdistribution
verification ladder T0 review T1 audit T2 compute T3 formal

Signed reviews

No signed human review yet.

0 comments
read the original abstract

We propose Future Discriminators for Generation (FUDGE), a flexible and modular method for controlled text generation. Given a pre-existing model G for generating text from a distribution of interest, FUDGE enables conditioning on a desired attribute a (for example, formality) while requiring access only to G's output logits. FUDGE learns an attribute predictor operating on a partial sequence, and uses this predictor's outputs to adjust G's original probabilities. We show that FUDGE models terms corresponding to a Bayesian decomposition of the conditional distribution of G given attribute a. Moreover, FUDGE can easily compose predictors for multiple desired attributes. We evaluate FUDGE on three tasks -- couplet completion in poetry, topic control in language generation, and formality change in machine translation -- and observe gains in all three tasks.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 12 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. ThinkRetrieve: Retrieval-Augmented Reasoning Traces for Test-Time Scaling

    cs.AI 2026-08 conditional novelty 6.0 of 10

    Per-step retrieval of solved exemplars injected into the reasoning trace improves test-time scaling accuracy, with up to 13.4 absolute points gained on AIME 2025.

  2. Artificial Epanorthosis: Why large language models overuse a classical rhetorical figure, and how to mitigate it

    cs.CL 2026-07 conditional novelty 6.0 of 10

    LLMs overuse the 'not X, but Y' self-correction pattern in persuasive registers and underuse it in informal Q&A; a prompt or a detachable LoRA dial adjusts it to human levels.

  3. Concurrent Image Understanding and Generation: Self-Correcting Coupled Markov Jump Processes

    cs.LG 2026-07 conditional novelty 6.0 of 10

    CO2Jump couples text and image denoising through cross-modal attention and remasking, achieving best joint accuracy on three concurrent-generation tasks.

  4. CALMA: A Process for Deriving Context-aligned Axes for Language Model Alignment

    cs.CY 2025-07 conditional novelty 6.0 of 10

    CALMA is a grounded-theory, participatory method for deriving community-specific language model alignment axes from open-ended user interactions and group discussion, piloted with two small groups.

  5. Aligning Frozen LLMs by Reinforcement Learning: An Iterative Reweight-then-Optimize Approach

    cs.LG 2025-06 conditional novelty 6.0 of 10

    IRO trains a sequence of small value functions that reweight a frozen LLM's own samples, achieving test-time alignment gains that outperform Best-of-N search on AlpacaEval 2.0 and TL;DR.

  6. Stream Aligner: Efficient Sentence-Level Alignment via Distribution Induction

    cs.CL 2025-01 reject novelty 6.0 of 10

    A sentence-level iterative correction method is claimed to improve large language model alignment and math reasoning, but the evidence is compromised by GPT-4 serving as both trainer and judge.

  7. Safeguarding Large Language Models in Real-time with Tunable Safety-Performance Trade-offs

    cs.CL 2025-01 conditional novelty 6.0 of 10

    SafeNudge uses a safety classifier on the model's own hidden states and a hidden nudge phrase to steer jailbroken LLM responses back to safe content, cutting unsafe responses from 55.4% to 25% on Llama-3.1-8B-Instruct.

  8. Time-Reversal Provides Unsupervised Feedback to LLMs

    cs.CL 2024-12 conditional novelty 6.0 of 10

    Time-reversed language models, which score queries given responses, improve best-of-N reranking, citation attribution, retrieval, and jailbreak filtering over forward perplexity baselines.

  9. Ontology-Constrained Generation of Domain-Specific Clinical Summaries

    cs.CL 2024-11 conditional novelty 6.0 of 10

    An ontology-guided constrained decoding method produces specialty-specific clinical summaries and lowers hallucination scores on MIMIC-III relative to greedy and beam search baselines.

  10. CoDe: Blockwise Control for Denoising Diffusion Models

    cs.CV 2025-02 conditional novelty 5.0 of 10

    CoDe applies blockwise best-of-N sampling during diffusion denoising, with Tweedie-based reward estimates, to align generated images to differentiable or non-differentiable rewards.

  11. Simple Guidance Mechanisms for Discrete Diffusion Models

    cs.LG 2024-12 conditional novelty 5.0 of 10

    Uniform-noise discrete diffusion trained with a continuous-time variational bound (UDLM) plus discrete classifier-free and classifier-based guidance improves controllable generation over autoregressive baselines on ge...

  12. COBART: Controlled, Optimized, Bidirectional and Auto-Regressive Transformer for Ad Headline Generation

    cs.CL 2026-07 conditional novelty 4.0 of 10

    Prefixing BART encoder input with bucketized CTR and length control tokens lets a single fine-tuned model generate ad headlines with controllable length and higher estimated CTR than prior baselines.

Pith tools