olly-styles/WorkBench issues and pull requests

#48 - Readme correction

Issue - State: closed - Opened by esbovee 10 months ago - 1 comment

#47 - Add README

Pull Request - State: closed - Opened by olly-styles 11 months ago

#46 - DATA0035 - Fix analytics eval bugs and update plots

Pull Request - State: closed - Opened by olly-styles 11 months ago

#45 - DATA0034 - Review mistakes in GPT4 from full run to find bugs

Pull Request - State: closed - Opened by olly-styles 11 months ago - 1 comment

#44 - DATA0033 - Filter results by all domains vs single domain

Pull Request - State: closed - Opened by olly-styles 11 months ago

#43 - add split of errors by actions vs no actions required in ground truth

Pull Request - State: closed - Opened by sammiller06 11 months ago

#42 - Flag off-by-one error for dates in predictions

Pull Request - State: closed - Opened by sammiller06 11 months ago

#41 - Update analytics data-gen and multi-domain bugs

Pull Request - State: closed - Opened by sammiller06 11 months ago

#40 - DATA0031 - Fix bugs found in full run (part 1)

Pull Request - State: closed - Opened by olly-styles 11 months ago

#39 - DATA0029 - Compare results of function-based-eval to compare with outcome-centric-eval

Pull Request - State: closed - Opened by olly-styles 11 months ago

#38 - Sam/multi domain templates

Pull Request - State: closed - Opened by sammiller06 11 months ago

#37 - DATA0028 - Bug fixes found on full run

Pull Request - State: closed - Opened by olly-styles 11 months ago

#36 - DATA0027 - Replace mistral-7b with a larger model

Pull Request - State: closed - Opened by olly-styles 11 months ago

#35 - DATA0026 - Multiple versions of wording per template

Pull Request - State: closed - Opened by olly-styles 11 months ago

#34 - DATA0025 - Make data generation more realistic

Pull Request - State: closed - Opened by olly-styles 11 months ago

#33 - DATA0024 - Generate a new query if it's a duplicate

Pull Request - State: closed - Opened by olly-styles 11 months ago

#32 - Analytics toolkit outputs series instead of point-in-time estimates

Pull Request - State: closed - Opened by sammiller06 11 months ago

#31 - DATA0023 - Plotting functions and better way to get results

Pull Request - State: closed - Opened by olly-styles 11 months ago

#30 - DATA0022 - Categorise agent errors

Pull Request - State: closed - Opened by olly-styles 12 months ago

#29 - Add capability to restrict tool selection to relevant domains

Pull Request - State: closed - Opened by sammiller06 12 months ago - 1 comment

#28 - DATA0022 - Categorise agent errors

Pull Request - State: closed - Opened by olly-styles 12 months ago - 1 comment

#27 - DATA0021 - Fix bug with lower vs upper case

Pull Request - State: closed - Opened by olly-styles 12 months ago

#26 - DATA020 - Fix single vs double quote issue

Pull Request - State: closed - Opened by olly-styles 12 months ago - 1 comment

#25 - Added query templates

Pull Request - State: closed - Opened by sammiller06 12 months ago

#24 - DATA019 - Implement CRM toolkit

Pull Request - State: closed - Opened by olly-styles 12 months ago

#23 - DATA018 - Implement project management toolkit

Pull Request - State: closed - Opened by olly-styles 12 months ago

#22 - DATA016 - Add more email query/answer templates

Pull Request - State: closed - Opened by olly-styles 12 months ago

#21 - DATA017 - Remove in_logic flag

Pull Request - State: closed - Opened by olly-styles 12 months ago

#20 - DATA015 - Bugs with multi-step calendar queries

Pull Request - State: closed - Opened by olly-styles 12 months ago - 1 comment

#19 - DATA014 - Add exception handler for when prompt exceeds context window length

Pull Request - State: closed - Opened by olly-styles almost 1 year ago

#18 - DATA013 - Test that we don't have any emails in the future

Pull Request - State: closed - Opened by olly-styles about 1 year ago

#17 - DATA012 - Evaluate a non-GPT model as part of the baseline

Pull Request - State: closed - Opened by olly-styles about 1 year ago

#16 - Refactor question template logic and add some templates

Pull Request - State: closed - Opened by sammiller06 about 1 year ago

#15 - DATA010 - Send/Forward email does not take any action

Pull Request - State: closed - Opened by olly-styles about 1 year ago

#14 - data009 - Re-add side effects to eval

Pull Request - State: closed - Opened by olly-styles about 1 year ago

#13 - data008 - Write more tests for the generated data

Pull Request - State: closed - Opened by olly-styles about 1 year ago

#12 - Eval based on state change

Pull Request - State: closed - Opened by sammiller06 about 1 year ago

#11 - Eval based on state change

Pull Request - State: closed - Opened by olly-styles about 1 year ago

#10 - data006 - Implement analytics toolkit

Pull Request - State: closed - Opened by olly-styles about 1 year ago - 1 comment

#9 - Fix evaluation of GPT4 calendar events

Pull Request - State: closed - Opened by olly-styles about 1 year ago - 1 comment

#8 - Store results more intelligently and evaluate multiple at once

Pull Request - State: closed - Opened by sammiller06 about 1 year ago

#7 - Answer generation script outputs accuracy

Pull Request - State: closed - Opened by sammiller06 about 1 year ago

#6 - Email and multi-domain questions. Results for draft 2 of paper

Pull Request - State: closed - Opened by olly-styles about 1 year ago

#5 - Fix toolkit naming

Pull Request - State: closed - Opened by sammiller06 about 1 year ago

#4 - Path fixes

Pull Request - State: closed - Opened by sammiller06 about 1 year ago

#3 - Make calendar questions more natural-sounding

Pull Request - State: closed - Opened by sammiller06 about 1 year ago

#2 - Check for side effects

Pull Request - State: closed - Opened by olly-styles about 1 year ago

#1 - Email and calendar tools

Pull Request - State: closed - Opened by olly-styles about 1 year ago - 1 comment

GitHub / olly-styles/WorkBench issues and pull requests