Kyle Richardson

yakazimir

·

https://www.krichardson.me/

yakazimir

AI & ML interests

NLP, ML

Recent Activity

authored a paper 1 day ago

Operads for compositional reasoning in LLMs

authored a paper 1 day ago

From Token Probabilities to Semantic Constraints: Towards Declarative Probabilistic Evaluation of Language Models

authored a paper 1 day ago

Operadic consistency: a label-free signal for compositional reasoning failures in LLMs

View all activity

Organizations

yakazimir 's datasets 14

yakazimir/on_policy_human_ai

Viewer • Updated Apr 12, 2025 • 92.5k • 80

yakazimir/preference_mixture

Viewer • Updated Feb 26, 2025 • 81.8k • 41

yakazimir/ultrafeedback_olmo1b_ref

Viewer • Updated Jan 14, 2025 • 62.5k • 39

yakazimir/mistral-instruct-ultrafeedback

Viewer • Updated Nov 14, 2024 • 62.7k • 19

yakazimir/ultrafeedback_binarized

Viewer • Updated Nov 14, 2024 • 63.1k • 15 • 1

yakazimir/llama3-ultrafeedback-armorm

Viewer • Updated Nov 14, 2024 • 61.8k • 40

yakazimir/preference_mix

Preview • Updated Sep 30, 2024 • 35

yakazimir/preference_alignment_ultra_cut

Viewer • Updated Sep 27, 2024 • 39.5k • 32

yakazimir/preference_tuning_hh_ultra

Viewer • Updated Sep 24, 2024 • 133k • 33 • 1

yakazimir/preference_alignment_total

Viewer • Updated Sep 22, 2024 • 153k • 24

yakazimir/preference_tuning_full

Updated Sep 22, 2024 • 5

yakazimir/preference_alignment_oasst

Viewer • Updated Sep 21, 2024 • 22.3k • 18

yakazimir/preference_tuning_hh

Viewer • Updated Sep 21, 2024 • 72.9k • 33 • 1

yakazimir/preference_tuning

Viewer • Updated Sep 19, 2024 • 63.1k • 9 • 1