lapisrocks/LanguageAgentTreeSearch

Issues

• Incorrect Simulation Loop Placement in MCTS Expansion Phase

  #35 opened 8 days ago by zzfoo
  0

• Is it correct to use the ground-truth answer on HotpotQA during MCTS?

  #6 opened a year ago by JianGuanTHU
  5

• Is the object of reflection for multiple nodes or a single node

  #34 opened 3 months ago by zulihit
  1

• Potential Bug in evaluate_node and question on self-consistency reward

  #33 opened 4 months ago by sjaelee25
  0

• Webshop is too slow

  #32 opened 4 months ago by SpaceLearner
  1

• code for “rap”, "react" and "tot" in the webshop environment

  #27 opened 5 months ago by George-Chia
  1

• Use of hidden test cases in programming tasks

  #28 opened 5 months ago by henryhungle
  1

• Inconsistent Accuracy Evaluation for GPT-3.5-Turbo Solutions

  #29 opened 5 months ago by VijayLingam95
  1

• Critical Bug: Incorrect Success Count in `programming/mcts.py` Severely Affects Reported Accuracy Metrics

  #30 opened 5 months ago by VijayLingam95
  1

• Request to release code for Game of 24 Ablations

  #31 opened 5 months ago by shubhamprshr27
  0

• #Hello, I don't understand how hotpotqa's rewards are calculated. The reward seems to be always 0. Can you explain how to assign rewards? Thank you!

  #26 opened 6 months ago by dailiokok
  0

• Hello, I don’t understand how hotpotqa’s reward is scored. It seems that when searching the wiki, the reward is not set to 1. What I always get is reward=0. Can you explain how to reward qa? Thank you.

  #20 opened 10 months ago by zouhuigang
  1

• up-to-date programming results

  #23 opened 6 months ago by jialin-song
  1

• LATS for local model

  #24 opened 6 months ago by cacayaya
  1

• About the number of mbpp data

  #25 opened 7 months ago by Qlalq
  0

• Please consider re-evaluating your HumanEval results.

  #22 opened 8 months ago by Honka1StarRail
  1

• prev_func_impl is not a list, can not be zipped

  #11 opened a year ago by shellhue
  1

• Error when running 'sh run_lats.sh' in programming directory

  #21 opened 10 months ago by HamedTaherkhani
  1

• Loop in the dfs_search execute only once

  #18 opened 10 months ago by xuanw7
  1

• Can we add JS capabilities to programming?

  #19 opened 10 months ago by aneeqahsan01
  1

• Question regarding Node Evaluation Phase

  #17 opened a year ago by andreasbinder
  3

• Programming trajectories

  #16 opened a year ago by jialin-song
  4

• Having trouble running lats on hotpot in 'standard' prompt sample mode

  #13 opened a year ago by dxing-cs
  2

• Lets package Lats for pypi

  #8 opened a year ago by maxsu
  2

• local running

  #14 opened a year ago by evolu8
  1

• `depth` variable isn't used in `run_lats` function

  #12 opened a year ago by knotgrass
  0

• LATS uses real labels in HumanEval for reward calculation. This seems to be quite different from reflexion (it only uses internal tests). Is this correct?

  #10 opened a year ago by geekan
  3

• Bug in LATs?

  #9 opened a year ago by exrhizo
  2

• The number of humaneval datasets is 160, not 164

  #5 opened a year ago by jiashenggu
  1

• code for the webshop is missing

  #1 opened a year ago by JosephCurwin
  2