Spaces:
Sleeping
Sleeping
Commit History
-
Fix JSON response parsing in ux_testing_fsm function to directly access the response object.
0e5d4ec -
added ux_testing api
e79e607
-
Enhance answer generation by adding optional context parameter to relevant functions and update .gitignore to include test directory.
633edb2
-
Clarify verbosity guidelines in prompts to avoid unnecessary elaboration.
2f7158d -
using litellm - for now shuffle disabled
4bb354c
-
Update requirements to include TinyTroupeFork repository
2194a0e
-
fix
a3fe22b -
trying with english
9b8a5e1 -
forced language of questions in prompts
1954e89 -
forced focus on questions in prompts
436adc8 -
temperature to 0.5 for answers
29ead85 -
now using 4.1mini with temperature 1
70f896f
-
fixed models and temp for prod
4de89c5
-
removed parameters from eval script
2ddfbfe -
fix on persona generation
355c747 -
removed tinytroup from requirements for simplicity
21379ad -
restored function for parameters generation - just for completeness
85aaefb -
removed parameters from generation personas api
e6ec255 -
added possibility to ask questions in bulk - now default
96205bb -
modified prompts for ask question and build personas - fixed temperature to 0 and model to 4.1 nano for testing
a7fab3f
-
modified persona generation logic: now fixed set of parameters
4934758 -
fix on answers
28d7d9e -
added function to save also users
f217abb -
prompts in a separated file
de56645 -
explicit parameters in call_llm function
61cbe0d
-
added variability in synth answers
379848c -
modify n of fleet in test script
d0d9c87 -
fixed n of parameters
9ba1483 -
modified gitignore
d2e51aa -
complete eval flow ready
5b9f598 -
using direct embedding comparison
598da24 -
added pairwise approach
516410c -
added eval script and data
1a9b41f -
fix in utils
b1f1aec
-
fix on split
7caf588 -
splitted users batch generation and answer generation
53f767a
-
added generate-users-answers endpoint
2c4400c -
increased to 20 params per user
5002a55
-
added language in prompt for chat with persona
ca7243d
-
fix prompt report
0493d58 -
all questions and users go in parallel when asking questions
1ae0452
-
fixed call_llm and models in apis
fcd6884 -
added test script
63a2fd9 -
cleaned code - improved response format - splitted generate users adding new api for parameters
066fafa
-
trying gpt nano 4.1
11a0ed1 -
refined prompts in dt flow
94249e4 -
increased max personas to generate
f3f4b31 -
now returning only interactions of the assistant in chat-with-twin
ce097c1 -
fix save person api
320ae86
-
modified reset agents
1afedc0