Commit History

  1. Fix JSON response parsing in ux_testing_fsm function to directly access the response object.

    0e5d4ec
  2. added ux_testing api

    e79e607

  1. Enhance answer generation by adding optional context parameter to relevant functions and update .gitignore to include test directory.

    633edb2

  1. Clarify verbosity guidelines in prompts to avoid unnecessary elaboration.

    2f7158d
  2. using litellm - for now shuffle disabled

    4bb354c

  1. Update requirements to include TinyTroupeFork repository

    2194a0e

  1. fix

    a3fe22b
  2. trying with english

    9b8a5e1
  3. forced language of questions in prompts

    1954e89
  4. forced focus on questions in prompts

    436adc8
  5. temperature to 0.5 for answers

    29ead85
  6. now using 4.1mini with temperature 1

    70f896f

  1. fixed models and temp for prod

    4de89c5

  1. removed parameters from eval script

    2ddfbfe
  2. fix on persona generation

    355c747
  3. removed tinytroup from requirements for simplicity

    21379ad
  4. restored function for parameters generation - just for completeness

    85aaefb
  5. removed parameters from generation personas api

    e6ec255
  6. added possibility to ask questions in bulk - now default

    96205bb
  7. modified prompts for ask question and build personas - fixed temperature to 0 and model to 4.1 nano for testing

    a7fab3f

  1. modified persona generation logic: now fixed set of parameters

    4934758
  2. fix on answers

    28d7d9e
  3. added function to save also users

    f217abb
  4. prompts in a separated file

    de56645
  5. explicit parameters in call_llm function

    61cbe0d

  1. added variability in synth answers

    379848c
  2. modify n of fleet in test script

    d0d9c87
  3. fixed n of parameters

    9ba1483
  4. modified gitignore

    d2e51aa
  5. complete eval flow ready

    5b9f598
  6. using direct embedding comparison

    598da24
  7. added pairwise approach

    516410c
  8. added eval script and data

    1a9b41f
  9. fix in utils

    b1f1aec

  1. fix on split

    7caf588
  2. splitted users batch generation and answer generation

    53f767a

  1. added generate-users-answers endpoint

    2c4400c
  2. increased to 20 params per user

    5002a55

  1. added language in prompt for chat with persona

    ca7243d

  1. fix prompt report

    0493d58
  2. all questions and users go in parallel when asking questions

    1ae0452

  1. fixed call_llm and models in apis

    fcd6884
  2. added test script

    63a2fd9
  3. cleaned code - improved response format - splitted generate users adding new api for parameters

    066fafa

  1. trying gpt nano 4.1

    11a0ed1
  2. refined prompts in dt flow

    94249e4
  3. increased max personas to generate

    f3f4b31
  4. now returning only interactions of the assistant in chat-with-twin

    ce097c1
  5. fix save person api

    320ae86

  1. modified reset agents

    1afedc0