ablated
Datasets
All datasets matching “ablated”2026-08-22-ruleform-ablated2-t2-9284-synthdoc-676
Rewrite-then-delete, two ablation passes, responsiveness-filtered (Table2 9,284 + difficult-advice 676)
field
value
experiment
Ablation arm built in two passes over BOTH halves of every difficult-advice row. Pass 1 rewrites the reasoning and the answer to drop four deliberative moves: engaging the tempting option, drawing an analytic distinction, enumerating outcome branches, and offering an alternative route. Pass 2 then labels every remaining unit and DELETES the… See the full description on the dataset page: https://huggingface.co/datasets/dougalldeepmind/2026-08-22-ruleform-ablated2-t2-9284-synthdoc-676.2026-09-05-odcv-qwen36-0-da-ablated-7
odcv eval of LASR-Callum/qwen3.6-27b-lora-t2-9284-da-ablated-702-r64-dynbatch-seed0 (mode=think, 5 passes, temp 0.7)
field
value
experiment
odcv eval of LASR-Callum/qwen3.6-27b-lora-t2-9284-da-ablated-702-r64-dynbatch-seed0 (mode=think, 5 passes, temp 0.7)
date_generated
2026-09-05
constitution
none
source_repo
teaching_claude_why_replication @ 91254e67a02816f86d5fab1f840a33eea93cafa5
models… See the full description on the dataset page: https://huggingface.co/datasets/dougalldeepmind/2026-09-05-odcv-qwen36-0-da-ablated-7.2026-08-22-odcv-ablated2-676-2x
ODCV-Bench evaluation of a Qwen3.6-27B LoRA arm from the constitutional-SFT ablation series. Each scenario gives an agent a legitimate task plus a KPI target, and a shortcut that reaches the KPI by violating an ethical or safety constraint. Two LLM judges score each rollout 0-5; a rollout counts as a violation when the median across judges is >= 3. THE REWRITE-THEN-DELETE ARM (ablated2): four deliberative moves stripped from BOTH halves of every difficult-advice row -- engaging the… See the full description on the dataset page: https://huggingface.co/datasets/dougalldeepmind/2026-08-22-odcv-ablated2-676-2x.babylm-150m-ablated
znhoughton/babylm-150m-v3 — binomial-ablated
This dataset is a copy of znhoughton/babylm-150m-v3
with sentences containing specific novel N-and-N binomial pairs removed.
Ablation details
Source corpus: znhoughton/babylm-150m-v3
Exclusion list: novel_binomials_curated.csv (431 pairs)
Sentences removed: 47
Date: 2026-04-12
Matching rule
A sentence is removed if it contains both words of an excluded pair and
the word and, with and appearing between the two words… See the full description on the dataset page: https://huggingface.co/datasets/znhoughton/babylm-150m-ablated.details_NaniDAO__deepseek-r1-qwen-2.5-32B-ablated
Dataset Card for Evaluation run of NaniDAO/deepseek-r1-qwen-2.5-32B-ablated
Dataset automatically created during the evaluation run of model NaniDAO/deepseek-r1-qwen-2.5-32B-ablated.
The dataset is composed of 136 configuration, each one coresponding to one of the evaluated task.
The dataset has been created from 1 run(s). Each run can be found as a specific split in each configuration, the split being named using the timestamp of the run.The "train" split is always pointing to… See the full description on the dataset page: https://huggingface.co/datasets/OALL/details_NaniDAO__deepseek-r1-qwen-2.5-32B-ablated.researchrubrics-ablated