How Value Induction Reshapes LLM Behaviour
Conversational Large Language Models are post-trained on language that expresses specific behavioural traits, such as curiosity, open-mindedness, and empathy, and values, such as helpfulness, harmlessness, and honesty. This is done to increase utility, ensure safety, and improve the experience of the people interacting with the model. However, values are complex and inter-related -- inducing one could modify behaviour on another. Further, inducing certain values can make models more addictive or sycophantic through language used in the generations, with a potential detrimental effect on the user. We investigate these and other unintended effects of value induction into models. We fine-tune models using curated value subsets of existing preference datasets, measuring the impact of value induction on expression of other values, model safety, anthropomorphic language, and various QA benchmarks. We find that (i) inducing values leads to expression of other related, and sometimes contrastive values, (ii) inducing positive values increases safety, and (iii) all values increase anthropomorphic language use, making models more validating and sycophantic.
Code (0)
등록된 구현이 없습니다.
Similar Papers 제목 키워드 기반
Through the Bottleneck: How Multi-head Latent Attention Separates Content from Position in Language Models
Multi-head Latent Attention (MLA), introduced in DeepSeek-V2, compresses key-value pairs through a shared low-rank bottleneck (cKV), achieving 81% KV-cache reduction during inference. Despite its adoption in massive prod…
Towards Fault Diagnosis in Induction Motor using Fractional Fourier Transform
A method for determining the current signature faults using Fractional Fourier Transform (FrFT) has been developed. The method has been applied to the real-time steady-state current of the inverter-fed high power inducti…
Fault DiagnosisInductive reasoning in humans and large language models
The impressive recent performance of large language models has led many to wonder to what extent they can serve as models of general intelligence or are similar to human cognition. We address this issue by applying GPT-3…
Emergent Regulatory Response and Shift of Half induction point under Resource Competition in Genetic circuits
Synthetic genetic circuits are implemented in living cells for their operation. During gene expression, proteins are produced from the respective genes, by formation of complexes through the process of transcription and …
TranslationMechanistic evaluation of Transformers and state space models
State space models (SSMs) for language modelling promise an efficient and performant alternative to quadratic-attention Transformers, yet show variable performance on recalling basic information from the context. While p…
Language ModellingMambaState Space Models