parameters()'s Component grouping is too fuzzy

m_lba <- readRDS(url("https://raw.github.com/DominiqueMakowski/cogmod/main/vignettes/models/m_lba1.rds"))
insight::find_parameters(m_lba)
#> Loading required namespace: rstan
#> $conditional
#> [1] "b_Intercept"      "b_ConditionSpeed"
#> 
#> $ndt
#> [1] "b_ndt_Intercept"      "b_ndt_ConditionSpeed"
#> 
#> $sigma
#> [1] "sigma"
#> 
#> $sigmabias
#> [1] "b_sigmabias_Intercept"      "b_sigmabias_ConditionSpeed"
#> 
#> $boundary
#> [1] "b_boundary_Intercept"      "b_boundary_ConditionSpeed"
insight::find_auxiliary(m_lba)
#> [1] "sigmabias" "boundary"  "ndt"       "sigma"

That model has one parameter type callled sigmabias, and is detected correctly by insight.

parameters::parameters(m_lba) |> 
  as.data.frame()
#>                    Parameter   Component     Median   CI      CI_low
#> 1                b_Intercept conditional  3.1557313 0.95   2.9303093
#> 2           b_ConditionSpeed conditional  0.5545620 0.95   0.2303213
#> 3            b_ndt_Intercept         ndt -1.4402565 0.95  -1.8081333
#> 4       b_ndt_ConditionSpeed         ndt -0.3078568 0.95  -0.6567503
#> 5                      sigma       sigma  1.0000000 0.95   1.0000000
#> 6      b_sigmabias_Intercept       sigma  0.5942473 0.95   0.2209075
#> 7 b_sigmabias_ConditionSpeed       sigma -1.1895320 0.95 -69.2162793
#> 8       b_boundary_Intercept    boundary  0.1223261 0.95  -0.3145768
#> 9  b_boundary_ConditionSpeed    boundary  0.5627345 0.95  -0.1123833
#>        CI_high    pd     Rhat ESS_tail
#> 1  3.419467085 1.000 1.027810      522
#> 2  0.901147971 1.000 1.008764      539
#> 3 -1.278263645 1.000 1.066640      124
#> 4 -0.007389358 0.978 1.177587      400
#> 5  1.000000000 1.000       NA       NA
#> 6  0.814467536 0.994 1.014741      411
#> 7 -0.561339126 1.000 1.686071       13
#> 8  0.762664058 0.681 1.062409      139
#> 9  1.306710988 0.940 1.237117      320

^{Created on 2026-08-19 with reprex v2.1.1}

But parameter lumps it together under Component = sigma, which is wrong, it is a separate component

Reproduce the issue with the m_lba example and compare insight::find_parameters() and parameters::parameters() output. Trace how the Component grouping is assigned, then verify that sigmabias is kept separate from sigma and that the displayed components match find_auxiliary().