Here's an example script you might consider trying (after amending it):

doc -e W // W is for each worksheet which is what you want to do
{
    %A=%H; // store active window name
    nlsf.init(); // initialize the fitter
    nlsf.func$=MyLine; // initialize the function to be used
    nlsf.p1=0.1; // initialize the parameter values
    nlsf.p2=0.1;
    nlsf.fitData$=%(%A,3); // initialize the dependent dataset
    layer -a; // rescale the graph when it comes up
    nlsf.wType=2; // statistical weighting
    nlsf.fit(100); // 100 iterations
    nlsf.xmode=0; // use same X as fit data
    %A!wks.col$(%A!wks.ncols).name$="FitData"; // rename fit data column to "FitData"
};

Note 1: Anything in red must be tailored to meet your needs.

1. Add nlsf.begin(); immediately after nlsf.init();.
2. Add nlsf.end(); immediately prior to the closing curly brace }.

In addition to this, you have two options:

1) You can continue to use nlsf.pastetoplot=1; in your script.

or

2) You can enable the Paste Parameters to Plot check box for that function (inside the fitter) and save the change to the function's definition file. Note: To save a function, select Function:Edit and click the Save button. Saving this change to the function will force the fitter to output its results to the plot for ALL future fits.

Finally, your best bet for viewing the parameters in a worksheet would be to click the Param. Worksheet button on the Results page (Action:Results) in the fitter. To duplicate this action through script, you must use the run.section() method to run the ParaConf section of the file called OFIT.OGS. In doing so, you must also pass the script the worksheet name. Furthermore, since the script loops through all worksheets, you must somehow prevent it from executing when it encounters these parameters worksheets. To do that, I have built in an if statement. See the fully edited version of the script below:

myinc=1;
doc -e W
{    
	%A=%H; // store active window name
	if(%[%A,11]!="Parameters")
	{
		%B=nlsf.func$;
		nlsf.init(); // initialize the fitter  
		nlsf.begin(); // complement to end()
		nlsf.func$=MyLine; // initialize function to be used    
		nlsf.p1=0.1; // initialize the parameter values    
		nlsf.p2=0.1;    
		nlsf.fitData$=%(%A,4); // initialize the dependent dataset    
		layer -a; // rescale the graph when it comes up    
		nlsf.wType=2; // statistical weighting    
		nlsf.fit(100); // 100 iterations    
		nlsf.xmode=0; // use same X as fit data
		run.section(fit.ogs,ParaConf,Parameters$(myinc));
		myinc++;
		nlsf.end(); // forces After Fit to run
	}
};