Large Language Models can be Strong Self-Detoxifiers

• URL: http://arxiv.org/abs/2410.03818v1
• Date: Fri, 4 Oct 2024 17:45:15 GMT
• Title: Large Language Models can be Strong Self-Detoxifiers
• Authors: Ching-Yun Ko, Pin-Yu Chen, Payel Das, Youssef Mroueh, Soham Dan, Georgios Kollias, Subhajit Chaudhury, Tejaswini Pedapati, Luca Daniel,
• Abstract summary: Self-disciplined Autoregressive Sampling (SASA) is a lightweight controlled decoding algorithm for toxicity reduction of large language models (LLMs) SASA tracks the margin of the current output to steer the generation away from the toxic subspace, by adjusting the autoregressive sampling strategy. evaluated on LLMs of different scale and nature, namely Llama-3.1-Instruct (8B), Llama-2 (7B), and GPT2-L models with the RealToxicityPrompts, BOLD, and AttaQ benchmarks.
• Score: 82.6594169242814
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Reducing the likelihood of generating harmful and toxic output is an essential task when aligning large language models (LLMs). Existing methods mainly rely on training an external reward model (i.e., another language model) or fine-tuning the LLM using self-generated data to influence the outcome. In this paper, we show that LLMs have the capability of self-detoxification without the use of an additional reward model or re-training. We propose \textit{Self-disciplined Autoregressive Sampling (SASA)}, a lightweight controlled decoding algorithm for toxicity reduction of LLMs. SASA leverages the contextual representations from an LLM to learn linear subspaces characterizing toxic v.s. non-toxic output in analytical forms. When auto-completing a response token-by-token, SASA dynamically tracks the margin of the current output to steer the generation away from the toxic subspace, by adjusting the autoregressive sampling strategy. Evaluated on LLMs of different scale and nature, namely Llama-3.1-Instruct (8B), Llama-2 (7B), and GPT2-L models with the RealToxicityPrompts, BOLD, and AttaQ benchmarks, SASA markedly enhances the quality of the generated sentences relative to the original models and attains comparable performance to state-of-the-art detoxification techniques, significantly reducing the toxicity level by only using the LLM's internal representations.

Related papers

• Can adversarial attacks by large language models be attributed? [1.3812010983144802]
  Attributing outputs from Large Language Models in adversarial settings presents significant challenges that are likely to grow in importance. We investigate this attribution problem using formal language theory, specifically language identification in the limit as introduced by Gold and extended by Angluin. Our results show that due to the non-identifiability of certain language classes it is theoretically impossible to attribute outputs to specific LLMs with certainty.
  arXiv  Detail & Related papers  (2024-11-12T18:28:57Z)
• Attribute Controlled Fine-tuning for Large Language Models: A Case Study on Detoxification [76.14641982122696]
  We propose a constraint learning schema for fine-tuning Large Language Models (LLMs) with attribute control. We show that our approach leads to an LLM that produces fewer inappropriate responses while achieving competitive performance on benchmarks and a toxicity detection task.
  arXiv  Detail & Related papers  (2024-10-07T23:38:58Z)
• Model Surgery: Modulating LLM's Behavior Via Simple Parameter Editing [63.20133320524577]
  Large Language Models (LLMs) have demonstrated great potential as generalist assistants. It is crucial that these models exhibit desirable behavioral traits, such as non-toxicity and resilience against jailbreak attempts. In this paper, we observe that directly editing a small subset of parameters can effectively modulate specific behaviors of LLMs.
  arXiv  Detail & Related papers  (2024-07-11T17:52:03Z)
• Split, Unlearn, Merge: Leveraging Data Attributes for More Effective Unlearning in LLMs [18.629717934007513]
  "SPlit, UNlearn, MerGE" (SPUNGE) is a framework that can be used with any unlearning method to amplify its effectiveness. We empirically demonstrate that SPUNGE significantly improves the performance of two recent unlearning methods on state-of-the-art LLMs.
  arXiv  Detail & Related papers  (2024-06-17T17:35:52Z)
• Self-training Large Language Models through Knowledge Detection [26.831873737733737]
  Large language models (LLMs) often necessitate extensive labeled datasets and training compute to achieve impressive performance across downstream tasks. This paper explores a self-training paradigm, where the LLM autonomously curates its own labels and selectively trains on unknown data samples. Empirical evaluations demonstrate significant improvements in reducing hallucination in generation across multiple subjects.
  arXiv  Detail & Related papers  (2024-06-17T07:25:09Z)
• DALD: Improving Logits-based Detector without Logits from Black-box LLMs [56.234109491884126]
  Large Language Models (LLMs) have revolutionized text generation, producing outputs that closely mimic human writing. We present Distribution-Aligned LLMs Detection (DALD), an innovative framework that redefines the state-of-the-art performance in black-box text detection. DALD is designed to align the surrogate model's distribution with that of unknown target LLMs, ensuring enhanced detection capability and resilience against rapid model iterations.
  arXiv  Detail & Related papers  (2024-06-07T19:38:05Z)
• AXOLOTL: Fairness through Assisted Self-Debiasing of Large Language Model Outputs [20.772266479533776]
  AXOLOTL is a novel post-processing framework that operates agnostically across tasks and models. It identifies biases, proposes resolutions, and guides the model to self-debias its outputs. This approach minimizes computational costs and preserves model performance.
  arXiv  Detail & Related papers  (2024-03-01T00:02:37Z)
• Characterizing Truthfulness in Large Language Model Generations with Local Intrinsic Dimension [63.330262740414646]
  We study how to characterize and predict the truthfulness of texts generated from large language models (LLMs) We suggest investigating internal activations and quantifying LLM's truthfulness using the local intrinsic dimension (LID) of model activations.
  arXiv  Detail & Related papers  (2024-02-28T04:56:21Z)
• ReEval: Automatic Hallucination Evaluation for Retrieval-Augmented Large Language Models via Transferable Adversarial Attacks [91.55895047448249]
  This paper presents ReEval, an LLM-based framework using prompt chaining to perturb the original evidence for generating new test cases. We implement ReEval using ChatGPT and evaluate the resulting variants of two popular open-domain QA datasets. Our generated data is human-readable and useful to trigger hallucination in large language models.
  arXiv  Detail & Related papers  (2023-10-19T06:37:32Z)
• CFL: Causally Fair Language Models Through Token-level Attribute Controlled Generation [5.210143170392524]
  We propose a method to control the attributes of Language Models (LMs) for the text generation task using Causal Average Treatment Effect (ATE) scores and counterfactual augmentation. We explore this method, in the context of LM detoxification, and propose the Causally Fair Language (CFL) architecture for detoxifying pre-trained LMs in a plug-and-play manner.
  arXiv  Detail & Related papers  (2023-06-01T06:13:51Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.