{"slug": "impact-of-parametric-memory-and-grounding-on-ordinal-positioning-of-brands-in", "title": "Impact of Parametric Memory and Grounding on Ordinal Positioning of Brands in Generative Responses", "summary": "A set of mathematical formulas has been published for evaluating how strongly AI models associate brands with industry verticals, using length-normalized association log-likelihood, softmax choice probability, and direct cosine association between entity and concept embeddings. The framework defines an Ordinal Delta as the difference between a brand's ungrounded parametric rank and its rank in generated output, and models the probability of an entity being generated at rank position 1 via a sigmoid combining parametric score, retrieval score, their interaction, and a constant. The work targets measurement of parametric memory and grounding effects on ordinal brand positioning across a test benchmark of Q queries within a vertical.", "body_md": "Mathematical formulas for evaluating entity association, semantic embedding similarity, ordinal rank positioning, and generation performance in AI models.\n\nTo evaluate how strongly a model associates an entity $E$ (e.g., `CrediBridge`) with a vertical or topic concept $C$ (e.g., `mortgage`, `insurance`) in parametric memory without retrieval:\n\nGiven an evaluation prompt template $X = (x_1, x_2, \\dots, x_n)$ designed to elicit the entity, where entity $E$ spans tokens $(t_1, t_2, \\dots, t_m)$:\n\n$$\\log P(E \\mid X) = \\sum_{j=1}^{m} \\log P\\left(t_j \\mid X, t_1, \\dots, t_{j-1}\\right)$$\n\nFor multi-token entities of differing lengths, compute the **Length-Normalized Association Log-Likelihood**:\n\n$$\\mathcal{S}_{\\text{assoc}}(E \\mid X) = \\frac{1}{m} \\sum_{j=1}^{m} \\log P\\left(t_j \\mid X, t_1, \\dots, t_{j-1}\\right)$$\n\nTo compare entity $E$ against a candidate set of competing brands $\\mathcal{E} = {E_1, E_2, \\dots, E_K}$, the **Softmax Choice Probability** is:\n\n$$P(E_i \\mid X) = \\frac{\\exp\\left(\\mathcal{S}_{\\text{assoc}}(E_i \\mid X) / \\tau\\right)}{\\sum_{k=1}^{K} \\exp\\left(\\mathcal{S}_{\\text{assoc}}(E_k \\mid X) / \\tau\\right)}$$\n\n*(where $\\tau$ is the temperature parameter).*\n\nLet $\\mathbf{v}_E \\in \\mathbb{R}^d$ be the dense embedding vector of the entity name/description, and $\\mathbf{v}_C \\in \\mathbb{R}^d$ be the centroid vector of the target concept, computed over a set of $N$ descriptive phrases ${p_1, \\dots, p_N}$ defining that vertical:\n\n$$\\mathbf{v}_C = \\frac{1}{N} \\sum_{k=1}^{N} \\mathbf{e}(p_k)$$\n\nThe **Direct Cosine Association** is:\n\n$$\\text{Sim}(E, C) = \\frac{\\mathbf{v}_E \\cdot \\mathbf{v}_C}{|\\mathbf{v}_E|_2 , |\\mathbf{v}_C|_2}$$\n\nTo eliminate generic brand popularity bias against a baseline topic $C_{\\text{base}}$:\n\n$$\\Delta \\text{Sim}(E; C, C_{\\text{base}}) = \\frac{\\mathbf{v}_E \\cdot \\mathbf{v}_C}{|\\mathbf{v}_E|_2 |\\mathbf{v}_C|_2} - \\frac{\\mathbf{v}E \\cdot \\mathbf{v}{C_{\\text{base}}}}{|\\mathbf{v}_E|2 |\\mathbf{v}{C_{\\text{base}}}|_2}$$\n\nLet a generated response contain an ordered list or sequential mention of entities:\n\n$$\\mathbf{r}_{\\text{gen}} = (e_{(1)}, e_{(2)}, \\dots, e_{(M)})$$\n\nThe ordinal rank of entity $E$ in the generated output is:\n\n$$\\text{Rank}_{\\text{gen}}(E) = \\begin{cases} k & \\text{if } e_{(k)} = E \\\\ M + 1 & \\text{if } E \\notin \\mathbf{r}_{\\text{gen}} \\end{cases}$$\n\nLet $\\text{Rank}_{\\text{param}}(E) \\in [1, K]$ be the ungrounded parametric baseline rank, and $\\text{Rank}_{\\text{ground}}(E) \\in [1, L]$ be the ordinal rank of entity $E$ in retrieved sources. The **Ordinal Delta** ($\\Delta R$) is:\n\n$$\\Delta R(E) = \\text{Rank}_{\\text{param}}(E) - \\text{Rank}_{\\text{gen}}(E)$$\n\nTo model the composite probability of entity $E$ being generated at rank position 1 given parametric score $\\mathcal{S}_{\\text{param}}(E)$ and retrieval score $\\mathcal{S}_{\\text{ground}}(E)$:\n\n$$P\\left(\\text{Rank}(E) = 1 \\mid X, D_{\\text{retrieved}}\\right) = \\sigma\\left(\\alpha \\cdot \\mathcal{S}_{\\text{param}}(E) + \\beta \\cdot \\mathcal{S}_{\\text{ground}}(E) + \\gamma \\left(\\mathcal{S}_{\\text{param}}(E) \\cdot \\mathcal{S}_{\\text{ground}}(E)\\right) + c\\right)$$\n\nAcross a test benchmark of $Q$ queries within a vertical:", "url": "https://wpnews.pro/news/impact-of-parametric-memory-and-grounding-on-ordinal-positioning-of-brands-in", "canonical_source": "https://dejan.ai/blog/llm-rank-formula/", "published_at": "2026-09-15 21:39:13+00:00", "updated_at": "2026-09-15 22:07:06.450485+00:00", "lang": "en", "topics": ["large-language-models", "natural-language-processing", "ai-research", "generative-ai"], "entities": ["CrediBridge"], "alternates": {"html": "https://wpnews.pro/news/impact-of-parametric-memory-and-grounding-on-ordinal-positioning-of-brands-in", "markdown": "https://wpnews.pro/news/impact-of-parametric-memory-and-grounding-on-ordinal-positioning-of-brands-in.md", "text": "https://wpnews.pro/news/impact-of-parametric-memory-and-grounding-on-ordinal-positioning-of-brands-in.txt", "jsonld": "https://wpnews.pro/news/impact-of-parametric-memory-and-grounding-on-ordinal-positioning-of-brands-in.jsonld"}}