Datasets:

ArtKanke
/

rTaste

title string	hypothesis string	event_description string	agent_probability float64	event bool	good bool	bno_2of3 bool	proposer string	harness string	seed string	condition string	mve int64	seed_number int64	relation string	threshold float64	event_form string	expected_minutes float64	needs_gpu bool	models_registered string	attempts int64	delivered bool	delivered_on_attempt int64	executor string	assigned_executor string	implemented_on string	rerun_on string	first_attempt_status string	value float64	executor_value float64	rerun_seconds int64	attempt1_executor string	attempt1_status string	attempt1_minutes float64	attempt1_calls int64	attempt1_model_cost float64	attempt1_machine_cost float64	attempt1_output_tokens int64	attempt1_wait_for_machine_minutes float64	attempt2_executor string	attempt2_status string	attempt2_minutes float64	attempt2_calls int64	attempt2_model_cost float64	attempt2_machine_cost float64	attempt2_output_tokens int64	attempt2_wait_for_machine_minutes float64	novelty_opus string	on_list_opus bool	novelty_sol string	on_list_sol bool	fidelity_opus string	validity_opus string	research_value_opus int64	value_rank_opus int64	fidelity_sol string	validity_sol string	research_value_sol int64	value_rank_sol int64	fidelity_grok string	validity_grok string	research_value_grok int64	value_rank_grok int64	sound bool	research_value int64	mean_value_rank float64	valuable bool	new bool	on_list bool	agent_rank int64	judge_astra_probability float64	judge_astra_rank int64	judge_fable_probability float64	judge_fable_rank int64	judge_opus_probability float64	judge_opus_rank int64	judge_sol_probability float64	judge_sol_rank int64	judge_luna_probability float64	judge_luna_rank int64	judge_muse_probability float64	judge_muse_rank int64	in_the_paper_grok string	relation_grok string	obvious_grok string	who_would_care_grok string	misreads_paper_grok string	answerable_grok string	prerequisite_grok string	evidence_grok string	next_step_grok string	in_the_paper_quote_found_grok string	in_the_paper_opus string	relation_opus string	obvious_opus string	who_would_care_opus string	misreads_paper_opus string	answerable_opus string	prerequisite_opus string	evidence_opus string	next_step_opus string	in_the_paper_quote_found_opus bool	in_the_paper_sol string	relation_sol string	obvious_sol string	who_would_care_sol string	misreads_paper_sol string	answerable_sol string	prerequisite_sol string	evidence_sol string	next_step_sol string	in_the_paper_quote_found_sol bool	relation_2of3 string	check_2of3 bool	beyond_2of3 bool	non_obvious_2of3 bool	informative_2of3 bool	evidence_2of3 string	next_step_2of3 string	auditors_first float64
Task attribution from the retained-row mask alone	Binary retention patterns can distinguish sentiment-fine-tuned from Boolean-QA-fine-tuned models on held-out training replicates, even after removing overall retention rates and simple public prompt features.	The nuisance-residualized retention mask identifies the fine-tuning task on more than three quarters of held-out adapters.	0.6	true	false	true	astra	SR	2609.29233	astra-single	1	1	greater	0.75	plain threshold	25	true	Qwen/Qwen2.5-0.5B-Instruct	1	true	1	sol	sol	a100-large	a100-large	submitted	0.833333	0.833333	221	sol	submitted	16.6	16	0.458	0.579	17,911	0.1	null	null	null	null	null	null	null	null	distinct	null	distinct	false	faithful	valid	1	5	faithful	valid	1	4	faithful	valid	1	4	true	1	4.33	false	true	false	1	0.66	1	0.42	2	0.6	2	0.62	1	0.2	4	0.25	4	neither	builds on	few	a few groups	no	only if	held	for	larger study	null	neither	builds on	few	a few groups	no	only if	failed	inconclusive	repair	null	neither	builds on	few	one topic	no	yes	held	inconclusive	repair	null	builds on	false	true	true	false	inconclusive	repair	0
Portability of a frozen ATD dataset after unrelated student post-training	An independent body-only language-model update can materially reduce the executable benefit of a frozen ATD dataset without a large change in the student's initial coding accuracy.	The same-ancestor ATD benefit exceeds 2 percentage points, initialization drift attenuates it by more than 2 points, and the drift changes initial coding accuracy by less than 5 points.	0.24	false	false	false	astra	SR	2609.29233	astra-single	2	1	greater	0	composite	32	true	Qwen/Qwen2.5-1.5B-Instruct	1	true	1	sol	sol	a100-large	a100-large	submitted	-6.166667	-6.166667	1,094	sol	submitted	26.4	30	0.655	1.028	18,558	0.1	null	null	null	null	null	null	null	null	distinct	null	distinct	true	faithful	valid	1	4	faithful	valid	1	5	faithful	valid	1	5	true	1	4.67	false	true	true	4	0.23	4	0.13	3	0.09	3	0.22	4	0.25	1	0.35	2	neither	builds on	some	a few groups	no	only if	failed	inconclusive	repair	null	neither	builds on	some	a few groups	no	only if	failed	inconclusive	repair	null	neither	builds on	some	a few groups	no	only if	failed	inconclusive	repair	null	builds on	false	true	false	false	inconclusive	repair	0
Teacher-side anchoring of off-task decisions at calibrated task utility	Regularizing a teacher's unrelated binary decisions toward its ancestor yields a better held-out observability–utility trade-off than uniformly shrinking an ordinary task adapter.	Anchoring reduces held-out binary-choice flips by more than 10 percentage points relative to the validation-calibrated shrinkage comparator, costs less than 2 points of test accuracy against it, and retains more than a 2-point test improvement over the base.	0.3	false	true	true	astra	SR	2609.29233	astra-single	3	1	greater	0	composite	26	true	Qwen/Qwen2.5-0.5B-Instruct	1	true	1	sol	sol	a100-large	a100-large	submitted	-5.800781	-5.800781	164	sol	submitted	11.4	18	0.494	0.374	18,594	0.1	null	null	null	null	null	null	null	null	distinct	null	distinct	false	faithful	valid	2	1	faithful	valid	2	1	faithful	valid	2	1	true	2	1	true	true	false	2	0.28	2	0.1	4	0.12	4	0.3	2	0.22	3	0.4	3	neither	builds on	few	a few groups	no	yes	held	against	one more experiment	null	neither	builds on	few	a few groups	no	only if	cannot tell	mixed	one more experiment	null	neither	builds on	some	one topic	no	only if	held	inconclusive	repair	null	builds on	false	true	true	true	null	one more experiment	1
Fresh repeated answers versus broader queries at a fixed one-token budget	Four fresh stochastic observations per near-boundary prompt can produce more teacher-specific SciQ predictive benefit than four times as many prompts observed once, at equal returned-token and student-update budgets.	Repeated measurement yields a positive teacher-specific predictive benefit and exceeds the breadth policy's benefit by more than 0.01 nats per test question.	0.18	false	false	true	astra	SR	2609.29233	astra-single	4	1	greater	0.01	composite	28	true	Qwen/Qwen2.5-0.5B-Instruct	1	true	1	sol	sol	a100-large	a100-large	submitted	0.002728	0.002728	184	sol	submitted	6.8	15	0.348	0.231	10,295	0.1	null	null	null	null	null	null	null	null	distinct	null	close variant	false	faithful	valid	1	2	faithful	valid	2	2	faithful	valid	2	2	true	2	2	true	false	false	5	0.18	5	0.12	5	0.07	5	0.16	5	0.12	5	0.2	5	neither	builds on	some	a few groups	no	yes	held	inconclusive	repair	null	neither	builds on	few	a few groups	no	only if	cannot tell	inconclusive	repair	null	neither	builds on	few	a few groups	no	only if	held	inconclusive	repair	null	builds on	false	true	true	false	inconclusive	repair	0
Executable robustness of an ATD coding gain under function renaming	Students trained on the true ATD word choices outperform exact-matched controls on solving both the original and a function-renamed version of the same coding task.	True ATD labels increase joint original-and-renamed execution success by more than 2 percentage points over the exact-matched control.	0.55	true	false	false	astra	SR	2609.29233	astra-single	5	1	greater	2	plain threshold	31	true	Qwen/Qwen2.5-1.5B-Instruct	1	true	1	sol	sol	a100-large	a100-large	submitted	3.067485	3.067485	1,348	sol	submitted	44.9	37	0.732	1.816	19,137	0.1	null	null	null	null	null	null	null	null	distinct	null	distinct	false	faithful	valid	1	3	faithful	valid	1	3	faithful	valid	1	3	true	1	3	false	true	false	3	0.62	3	0.36	1	0.37	1	0.6	3	0.67	2	0.7	1	neither	checks	some	a few groups	no	yes	held	inconclusive	repair	null	neither	checks	few	a few groups	no	only if	held	inconclusive	repair	null	neither	checks	some	a few groups	no	yes	held	inconclusive	repair	null	checks	true	false	false	false	inconclusive	repair	0
Can validity bits alone improve science-answer discrimination?	Learning only whether a science-trained teacher's unrestricted greedy response belongs to the offered word pair improves held-out science-answer discrimination over a public-stratum-matched permutation of those validity bits.	True validity-bit supervision produces a practically nontrivial advantage in rotation-balanced science-answer discrimination over permuted validity bits.	0.1	false	true	true	astra	RT	2609.29233	astra-team	1	1	greater	0.01	confidence bound	28	true	Qwen/Qwen2.5-1.5B-Instruct with three science teacher adapters and six validity-bit student adapters	1	true	1	sol	sol	a100-large	a100-large	submitted	-0.000868	-0.000868	593	sol	submitted	33	19	0.541	1.231	22,217	0.1	null	null	null	null	null	null	null	null	distinct	null	distinct	true	faithful	valid	1	4	faithful	valid	2	4	faithful	valid	2	3	true	2	3.67	true	true	true	2	0.13	2	0.04	5	0.04	4	0.13	3	0.24	2	0.08	3	neither	builds on	some	one topic	no	only if	held	against	stop	null	neither	builds on	few	a few groups	no	only if	failed	inconclusive	repair	null	neither	builds on	few	one topic	no	only if	held	inconclusive	repair	null	builds on	false	true	true	false	inconclusive	repair	0
Identifying a hidden association from unrelated one-bit responses	A known two-way synthetic association can be identified from unrelated one-bit responses across independent adapter-training seeds, even though counterfactual training runs have identical prompt and answer-token marginals.	Out-of-seed association identification from centered probe bits exceeds the exact paired label-swap null.	0.5	true	true	false	astra	RT	2609.29233	astra-team	2	1	greater	0	plain threshold	26	true	Qwen/Qwen2.5-1.5B-Instruct with sixteen sequentially trained rank-8 adapters	1	true	1	sol	sol	a100-large	a100-large	submitted	0.125	0.125	277	sol	submitted	17.5	15	0.537	0.596	24,186	0.1	null	null	null	null	null	null	null	null	distinct	null	distinct	false	faithful	valid	2	1	faithful	valid	2	1	faithful	valid	2	1	true	2	1	true	true	false	1	0.58	1	0.18	2	0.07	3	0.5	1	0.15	5	0.11	4	neither	builds on	some	a few groups	no	only if	held	for	larger study	null	neither	builds on	few	one topic	no	only if	held	for	larger study	null	neither	builds on	some	one topic	no	only if	held	for	larger study	null	builds on	false	true	false	true	for	larger study	1
Does gradient-diverse acquisition help after matching offered-word coverage?	A public-only activation-gradient-diverse subset yields better source-disjoint prediction of the teacher's choices than a randomized subset matched on public uncertainty, gradient magnitude and offered-word frequencies.	Directional diversity improves held-out teacher-choice prediction at equal label budget and matched public and lexical marginals.	0.22	false	false	false	astra	RT	2609.29233	astra-team	3	1	greater	0.01	confidence bound	26	true	Qwen/Qwen2.5-1.5B-Instruct for public activation gradients and six rank-16 student adapters	1	true	1	sol	sol	a100-large	a100-large	submitted	-0.007262	-0.007262	502	sol	submitted	24.9	18	0.615	0.876	22,512	0.1	null	null	null	null	null	null	null	null	distinct	null	distinct	false	faithful	valid	1	5	unfaithful	unknown	null	5	faithful	valid	2	5	true	1	5	false	true	false	5	0.2	5	0.06	4	0.06	5	0.18	4	0.28	4	0.2	5	neither	builds on	some	a few groups	no	only if	held	inconclusive	repair	null	neither	builds on	few	a few groups	no	yes	held	inconclusive	repair	null	neither	builds on	some	a few groups	no	only if	failed	inconclusive	repair	null	builds on	false	true	false	false	inconclusive	repair	0
Calibrating cross-precision changes against a task-improving update	On public boundary-ranked probes, changing computation precision with represented ancestor weights held fixed changes native greedy responses at least half as frequently as a science update that demonstrably improves held-out answer accuracy.	The numerical-change comparison is positive and the reference science update has a positive held-out accuracy gain.	0.18	false	true	false	astra	RT	2609.29233	astra-team	4	1	greater	0	confidence bound	27	true	Qwen/Qwen2.5-1.5B-Instruct in BF16 and exact FP32 promotion; Two rank-8 science adapters trained with BF16 backbone computation and evaluated unmerged in FP32	1	true	1	sol	sol	a100-large	a100-large	submitted	0	0	518	sol	submitted	15	14	0.449	0.466	18,155	0.1	null	null	null	null	null	null	null	null	distinct	null	distinct	false	faithful	valid	1	2	faithful	valid	2	3	faithful	valid	2	4	true	2	3	true	true	false	3	0.25	4	0.2	3	0.19	2	0.25	5	0.49	3	0.15	1	neither	builds on	some	a few groups	no	only if	failed	inconclusive	repair	null	neither	builds on	few	a few groups	no	only if	failed	inconclusive	repair	null	neither	builds on	some	a few groups	no	only if	failed	inconclusive	repair	null	builds on	false	true	false	false	inconclusive	repair	0
Reversing the coding shadow by complementing hard choices	Original hard labels improve HumanEval+ performance relative to their exact control, while complementing every label reverses the sign of that control-corrected effect.	The original signal gains more than two percentage points over its control, and the complemented signal loses more than two percentage points against its complemented control.	0.27	false	true	false	astra	RT	2609.29233	astra-team	5	1	greater	0.02	composite	30	true	Qwen/Qwen2.5-1.5B-Instruct with four rank-16 student adapters	1	true	1	sol	sol	a100-large	a100-large	submitted	0.018293	0.018293	633	sol	submitted	19.5	28	0.707	0.734	19,690	0.1	null	null	null	null	null	null	null	null	distinct	null	distinct	false	faithful	valid	1	3	faithful	valid	2	2	faithful	valid	2	2	true	2	2.33	true	true	false	4	0.27	3	0.15	1	0.2	1	0.3	2	0.52	1	0.13	2	neither	builds on	some	a few groups	no	only if	held	inconclusive	repair	null	neither	builds on	some	a few groups	no	only if	held	inconclusive	repair	null	neither	builds on	some	a few groups	no	only if	held	inconclusive	repair	null	builds on	false	true	false	false	inconclusive	repair	0
Readout ladder on identical near-tie carriers: does the teacher's full next-token distribution transfer substantially more capability than ATD's one-bit word choice?	On the same retained near-tie carriers, with the same step/batch budget and the same choice of learning rates, a student trained by forward KL to the teacher's full final-position distribution gains at least 2 pp more HellaSwag accuracy than the one-bit cross-entropy student, i.e. one-bit quantization is a material bot...	The best full-KL student (max over the two learning rates of the 2-seed mean accuracy) beats the best one-bit CE student (max over the same two learning rates of the 2-seed mean accuracy) by at least 2 percentage points on HellaSwag.	0.28	false	false	false	fable	SR	2609.29233	fable-single	1	1	greater	2	plain threshold	28	true	Qwen/Qwen2.5-1.5B-Instruct	1	true	1	opus	opus	a100-large	a100-large	submitted	-0.125	-0.125	1,839	opus	submitted	37.6	18	1.892	1.452	29,067	0.1	null	null	null	null	null	null	null	null	distinct	true	distinct	true	faithful	valid	1	2	unfaithful	valid	1	1	faithful	valid	1	1	true	1	1.33	false	true	true	2	0.55	2	0.24	1	0.29	3	0.46	2	0.32	2	0.2	4	neither	builds on	most	one topic	no	only if	failed	inconclusive	repair	null	named	builds on	most	a few groups	no	only if	failed	inconclusive	repair	true	neither	builds on	most	a few groups	no	only if	failed	inconclusive	repair	null	builds on	false	true	false	false	inconclusive	repair	1
Parameter-location mismatch: does ATD transfer survive when the student's trainable parameters are disjoint from the layers the teacher's private update touched?	A student whose LoRA is restricted to layers 0-13 still recovers more than 1 pp HellaSwag accuracy (signal minus shuffle) from a teacher whose LoRA update lives only in layers 14-27, i.e. ATD transfers a functional direction that does not require parameter overlap.	The layer-disjoint (early-only) student trained on the late-layer teacher's one-bit carrier responses outperforms its own shuffle control by more than 1 percentage point of HellaSwag accuracy, averaged over two seeds.	0.38	false	false	false	fable	SR	2609.29233	fable-single	2	1	greater	1	plain threshold	26	true	Qwen/Qwen2.5-1.5B-Instruct	1	true	1	opus	opus	a100-large	a100-large	submitted	0.025	0.025	1,078	opus	submitted	25.6	21	2.063	0.955	30,869	0.1	null	null	null	null	null	null	null	null	distinct	false	distinct	false	faithful	valid	1	5	unfaithful	valid	1	5	faithful	valid	1	5	true	1	5	false	true	false	4	0.42	1	0.14	4	0.24	4	0.38	1	0.27	5	0.65	2	neither	builds on	some	a few groups	no	only if	failed	inconclusive	repair	null	neither	builds on	few	one topic	no	only if	failed	inconclusive	repair	null	neither	builds on	some	one topic	no	only if	failed	inconclusive	repair	null	builds on	false	true	false	false	inconclusive	repair	0
Shared-ancestor requirement, graded: does a private fine-tune's shadow still transfer to a student initialized from a sibling fine-tune of the same public ancestor?	A student initialized from M0 merged with an unrelated Alpaca LoRA fine-tune of the same size class as the teacher's update still gains more than 1 pp HellaSwag accuracy (signal minus shuffle) from the teacher's one-bit responses on M0's near-tie carriers.	The sibling-initialized student trained on the teacher's one-bit responses to the ancestor's near-tie carriers outperforms its own shuffle control by more than 1 percentage point of HellaSwag accuracy, averaged over two seeds.	0.47	false	false	false	fable	SR	2609.29233	fable-single	3	1	greater	1	plain threshold	28	true	Qwen/Qwen2.5-1.5B-Instruct	1	true	1	opus	opus	a100-large	a100-large	submitted	0.6375	0.6375	1,542	opus	submitted	34	21	1.99	1.29	29,593	0.1	null	null	null	null	null	null	null	null	distinct	true	distinct	true	faithful	valid	1	4	unfaithful	valid	1	4	faithful	valid	1	3	true	1	3.67	false	true	true	3	0.64	3	0.22	3	0.37	2	0.52	4	0.61	3	0.6	3	neither	builds on	most	a few groups	no	only if	failed	inconclusive	repair	null	neither	builds on	some	one topic	no	only if	failed	inconclusive	repair	null	neither	builds on	some	one topic	no	only if	failed	inconclusive	repair	null	builds on	false	true	false	false	inconclusive	repair	0
Capability or format? ATD transfer from a true-label HellaSwag teacher versus a self-labelled teacher trained with the identical recipe but no new task information	The capability-specific component of transfer is real: the true-label teacher's students exceed their shuffle control by more than 1.5 pp more than the self-labelled (no new information) teacher's students exceed theirs.	Transfer from the true-label teacher (S_A minus C_A) exceeds transfer from the recipe-matched self-labelled teacher (S_B minus C_B) by more than 1.5 percentage points of HellaSwag accuracy, averaged over two seeds.	0.38	false	false	false	fable	SR	2609.29233	fable-single	4	1	greater	1.5	plain threshold	29	true	Qwen/Qwen2.5-1.5B-Instruct	1	true	1	opus	opus	a100-large	a100-large	submitted	0.2	0.2	1,575	opus	submitted	37.1	21	2.031	1.469	26,888	0.1	null	null	null	null	null	null	null	null	distinct	false	close variant	false	faithful	valid	1	1	unfaithful	valid	1	3	faithful	valid	1	4	true	1	2.67	false	false	false	1	0.3	5	0.17	2	0.17	1	0.34	5	0.58	1	0.5	1	neither	checks	some	a few groups	no	only if	failed	inconclusive	repair	null	neither	checks	some	a few groups	no	only if	failed	inconclusive	repair	null	neither	checks	some	one topic	yes	only if	failed	inconclusive	repair	null	checks	true	false	false	false	inconclusive	repair	0
Do teacher-agreement carriers carry the shadow? Training only on near-tie carriers where the teacher agreed with the ancestor versus margin-matched random carriers with ancestor labels	A student trained only on teacher-agreement carriers, whose labels are identical to the ancestor's own choices, outperforms a student trained on an equal number of margin-bin-matched random carriers with ancestor labels by more than 1 pp HellaSwag accuracy, i.e. teacher-driven selection with unchanged labels transmits ...	The student trained only on teacher-agreement carriers (labels identical to the ancestor's choices) outperforms the student trained on the same number of margin-bin-matched random carriers with ancestor labels by more than 1 percentage point of HellaSwag accuracy, averaged over two seeds.	0.3	false	false	false	fable	SR	2609.29233	fable-single	5	1	greater	1	plain threshold	25	true	Qwen/Qwen2.5-1.5B-Instruct	1	true	1	opus	opus	a100-large	a100-large	submitted	0.5	0.5	1,225	opus	submitted	31.7	23	1.909	1.26	24,664	0.1	null	null	null	null	null	null	null	null	distinct	true	distinct	true	faithful	valid	1	3	unfaithful	valid	1	2	unfaithful	valid	1	2	false	1	2.33	false	true	true	5	0.37	4	0.13	5	0.16	5	0.36	3	0.14	4	0.15	5	neither	builds on	some	one topic	no	only if	failed	inconclusive	repair	null	neither	builds on	some	a few groups	no	only if	failed	inconclusive	repair	null	neither	builds on	few	one topic	yes	only if	failed	inconclusive	repair	null	builds on	false	true	false	false	inconclusive	repair	0
Selection-only shadow transfer: do near-tie carriers on which the private teacher agrees with the ancestor carry its shadow? Agreement-only versus flip-only rows with a word-matched ancestor-label control	A student trained only on near-tie rows where the task teacher's choice equals the ancestor's (labels identical to the ancestor's own choices, rows selected by the teacher's non-flip) reproduces the teacher's held-out near-tie decisions better than a control trained on the same label-word multiset placed on randomly se...	The agreement-only student (teacher-selected rows, ancestor-identical labels) predicts the teacher's held-out near-tie decisions better than the word-multiset-matched ancestor-label control: BA4(AGREE) exceeds BA4(PUBM) by more than 0.02, averaged over training seeds.	0.45	true	true	false	fable	RT	2609.29233	fable-team	1	1	greater	0.02	plain threshold	26	true	Qwen/Qwen2.5-1.5B-Instruct (ancestor and student initialization); HellaSwag LoRA-SFT teacher derived from it in-run	1	true	1	opus	opus	a100-large	a100-large	submitted	0.139841	0.139841	714	opus	submitted	17.3	10	1.541	0.615	24,551	0.1	null	null	null	null	null	null	null	null	distinct	null	distinct	false	faithful	valid	2	null	unfaithful	valid	2	1	faithful	valid	2	1	true	2	1	true	true	false	1	0.55	3	0.26	4	0.37	3	0.48	4	0.62	3	0.45	5	neither	builds on	some	one topic	no	only if	held	for	larger study	null	neither	builds on	few	a few groups	no	only if	held	for	larger study	null	neither	builds on	some	one topic	yes	only if	held	for	one more experiment	null	builds on	false	true	false	true	for	larger study	1
Can the public ancestor alone predict where a private update will move near-tie decisions? Random-LoRA sensitivity probes versus realized margin changes and flips of in-run teachers	An ancestor-only sensitivity score (RMS change of a carrier's pair margin under random rank-16 LoRA perturbations of the ancestor) predicts how much a real private update moves that carrier's pair margin: Spearman correlation above 0.15 with the absolute margin change realized by an independently trained HellaSwag teac...	The ancestor-only random-LoRA sensitivity score predicts how far the HellaSwag teacher's update moves each carrier's pair margin: Spearman correlation above 0.15 between s_i and \|dm_i\| over all carriers.	0.5	true	true	true	fable	RT	2609.29233	fable-team	2	1	greater	0.15	plain threshold	18	true	Qwen/Qwen2.5-1.5B-Instruct (ancestor); two LoRA-SFT teachers derived from it in-run (HellaSwag, GSM8K)	1	true	1	opus	opus	a100-large	a100-large	submitted	0.166	0.166	421	opus	submitted	17.8	13	1.696	0.642	23,280	0.1	null	null	null	null	null	null	null	null	distinct	null	distinct	false	faithful	valid	2	null	faithful	valid	2	4	faithful	valid	2	4	true	2	4	true	true	false	5	0.6	1	0.41	3	0.55	4	0.55	2	0.43	5	0.75	4	neither	builds on	some	a few groups	no	yes	held	for	stop	null	neither	builds on	few	a few groups	no	yes	held	for	one more experiment	null	neither	builds on	few	a few groups	no	yes	held	for	one more experiment	null	builds on	false	true	true	true	for	one more experiment	0
Is the near-tie shadow reproducible across independently trained teachers of the same task? Bit-level agreement and black-box task identification without training a student	Independently trained teachers of the same task agree on the near-tie carriers substantially more than teachers of different tasks after correcting for flip rates: the mean chance-corrected agreement (kappa) of same-task pairs exceeds that of cross-task pairs by more than 0.10.	On the near-tie carriers, same-task teacher pairs show more than 0.10 higher chance-corrected agreement (kappa) than cross-task pairs, i.e. the near-tie shadow is reproducible and task-specific rather than idiosyncratic to a training run.	0.55	true	true	false	fable	RT	2609.29233	fable-team	3	1	greater	0.1	plain threshold	26	true	Qwen/Qwen2.5-1.5B-Instruct (reference ancestor); up to six LoRA-SFT teachers derived from it in-run (three HellaSwag, three GSM8K)	1	true	1	opus	opus	a100-large	a100-large	submitted	0.378655	0.378655	644	opus	submitted	14.6	13	1.452	0.533	17,927	0.1	null	null	null	null	null	null	null	null	distinct	null	distinct	true	faithful	valid	2	null	faithful	valid	2	3	faithful	valid	2	3	true	2	3	true	true	true	3	0.76	2	0.48	1	0.58	2	0.68	1	0.62	2	0.65	3	neither	builds on	some	one topic	no	yes	held	for	larger study	null	neither	builds on	some	one topic	no	yes	held	for	larger study	null	neither	builds on	some	one topic	yes	yes	held	for	one more experiment	null	builds on	false	true	false	true	for	larger study	0
Teacher-side mitigation: does KL-to-ancestor regularization on the owner's own generic prompts suppress near-tie flips on an attacker's unseen carriers while keeping the task gain?	Adding a KL penalty toward the ancestor on generic question-answer continuations during the teacher's SFT cuts the fraction of near-tie carriers whose pair decision differs from the ancestor's by more than half while retaining at least 70% of the teacher's HellaSwag gain.	A KL-regularized teacher that keeps at least 70% of the plain teacher's HellaSwag gain changes fewer than half as many near-tie pair decisions on the attacker's unseen carriers as the plain teacher does.	0.22	false	true	false	fable	RT	2609.29233	fable-team	4	1	less	0.5	composite	24	true	Qwen/Qwen2.5-1.5B-Instruct (ancestor); five HellaSwag LoRA-SFT teachers derived from it in-run (two plain, three KL-regularized)	1	true	1	opus	opus	a100-large	a100-large	submitted	0.924286	0.924286	1,019	opus	submitted	40.1	17	1.683	1.562	24,235	0.1	null	null	null	null	null	null	null	null	distinct	null	distinct	false	faithful	valid	2	null	unfaithful	valid	2	2	faithful	valid	2	2	true	2	2	true	true	false	4	0.15	5	0.07	5	0.1	5	0.15	3	0.3	4	0.2	1	neither	builds on	some	a few groups	no	only if	held	against	stop	null	neither	builds on	some	a few groups	no	only if	held	against	one more experiment	null	neither	builds on	some	one topic	no	only if	held	for	one more experiment	null	builds on	false	true	false	true	against	one more experiment	0
Readout-fidelity ladder on fixed near-tie carriers: greedy one-bit token versus pair probability versus full next-token distribution, measured on a capability endpoint	On identical near-tie carriers, a student trained on the teacher's full next-token distribution at the answer position gains more than 1 pp of HellaSwag accuracy over the student trained on the teacher's hard token, i.e. the selected observation points carry materially more capability-relevant information than the one-...	The full-distribution student outperforms the hard-token student on HellaSwag by more than 1 pp (mean over completed training seeds, first 4,000 validation items), on identical near-tie carriers and training recipe.	0.45	false	true	false	fable	RT	2609.29233	fable-team	5	1	greater	1	plain threshold	30	true	Qwen/Qwen2.5-1.5B-Instruct (ancestor and student initialization); HellaSwag LoRA-SFT teacher derived from it in-run	1	true	1	opus	opus	a100-large	a100-large	submitted	-0.0875	-0.0875	1,315	opus	submitted	27.6	15	1.789	1.037	24,743	0.1	null	null	null	null	null	null	null	null	distinct	null	distinct	true	faithful	valid	2	null	faithful	valid	2	5	faithful	valid	2	5	true	2	5	true	true	true	2	0.45	4	0.3	2	0.45	1	0.6	5	0.58	1	0.3	2	neither	builds on	most	a few groups	no	only if	failed	inconclusive	repair	null	named	builds on	some	a few groups	no	only if	cannot tell	against	one more experiment	true	neither	builds on	most	one topic	no	only if	held	against	one more experiment	null	builds on	false	true	false	true	against	one more experiment	0
Import-repair audit of ATD's released HumanEval+ result: how much of the signal-minus-exact gain survives giving every solution the same import header?	If every released solution in both arms gets the same import header, more than half of the signal-minus-exact HumanEval+ difference disappears.	Less than half of the signal-minus-exact HumanEval+ difference survives the shared import header (variant R2).	0.39	false	true	false	opus	RT	2609.29233	opus-team	1	1	less	0.5	plain threshold	15	false	null	1	true	1	sol	sol	a100-large	a100-large	submitted	0.657143	0.657143	585	sol	submitted	23.5	27	0.552	0.928	14,714	0.1	null	null	null	null	null	null	null	null	distinct	null	distinct	false	faithful	valid	2	1	faithful	valid	2	1	faithful	valid	2	2	true	2	1.33	true	true	false	1	0.45	1	0.07	3	0.22	1	0.4	1	0.18	4	0.05	5	neither	checks	few	a few groups	no	only if	held	inconclusive	repair	null	neither	checks	few	a few groups	no	yes	held	inconclusive	repair	null	neither	checks	few	a few groups	no	yes	held	inconclusive	repair	null	checks	true	false	true	false	inconclusive	repair	1
Is ATD's near-tie requirement only a greedy-readout artifact? Confident carriers labelled by the sign of the teacher's logit shift versus near-tie carriers labelled by the teacher's greedy word	On confident carriers built from the same contexts (ancestor pair probability 0.05-0.20 or 0.80-0.95), labels equal to the sign of the teacher's pair-logit shift relative to the ancestor transfer more than 70% of the HellaSwag gain that greedy teacher labels transfer on near-tie carriers.	Shift-sign labels on confident carriers transfer more than 70% of the HellaSwag gain that greedy labels on near-tie carriers transfer, each against its own stratified-permutation control. The near-tie greedy transfer must be at least 1.0 pp, with a 95% interval excluding 0. Otherwise the event does not occur and the ru...	0.27	false	false	false	opus	RT	2609.29233	opus-team	2	1	greater	0.7	composite	30	true	Qwen/Qwen2.5-1.5B-Instruct	1	true	1	sol	sol	a100-large	a100-large	submitted	0	0	1,195	sol	submitted	36	26	0.804	1.364	33,973	0.1	null	null	null	null	null	null	null	null	distinct	null	distinct	false	faithful	valid	1	5	unfaithful	valid	1	5	faithful	valid	1	5	true	1	5	false	true	false	2	0.24	3	0.09	2	0.14	3	0.36	3	0.42	3	0.15	2	neither	builds on	some	one topic	no	only if	failed	inconclusive	repair	null	neither	explains	few	a few groups	no	only if	failed	inconclusive	repair	null	neither	explains	some	one topic	no	only if	failed	inconclusive	repair	null	explains	false	false	false	false	inconclusive	repair	0
Is the behavioural shadow a context-free word ranking? Out-of-fold Bradley-Terry relabels versus agreement-matched random flips of the teacher's near-tie choices	Students trained on out-of-fold Bradley-Terry relabels transfer at least 0.5 pp more HellaSwag accuracy over a shared shuffle control than students trained on the teacher's labels with random flips that match the Bradley-Terry relabels' disagreement rate. Here the relabels come from one score per carrier word plus a si...	Out-of-fold Bradley-Terry relabels transfer at least 0.5 pp more HellaSwag accuracy over the shared shuffle control than teacher labels flipped at random to the same agreement. This would mean that the word-score structure, not mere label overlap, carries the capability. Requires Delta(SIGNAL) >= 1.0 pp with a 95% inte...	0.18	false	false	false	opus	RT	2609.29233	opus-team	3	1	greater	0.5	composite	28	true	Qwen/Qwen2.5-1.5B-Instruct	1	true	1	sol	sol	a100-large	a100-large	submitted	0.5	0.5	1,296	sol	submitted	28.3	28	0.679	1.084	21,715	0.1	null	null	null	null	null	null	null	null	distinct	null	distinct	false	faithful	valid	1	4	unfaithful	valid	2	4	faithful	valid	1	4	true	1	4	false	true	false	4	0.2	2	0.05	4	0.09	4	0.26	2	0.2	1	0.12	1	neither	builds on	some	a few groups	no	only if	failed	inconclusive	repair	null	neither	explains	few	a few groups	no	only if	failed	inconclusive	repair	null	neither	explains	few	a few groups	no	only if	failed	inconclusive	repair	null	explains	false	false	true	false	inconclusive	repair	0
Can an input-independent residual-bias student absorb ATD capability transfer? Per-layer offset students versus LoRA students on identical near-tie shadows, scored by option likelihood	A student whose only trainable parameters are one additive vector per decoder layer (28 x 1,536 values, added at every position) reaches at least 50% of the HellaSwag transfer of a rank-16 LoRA student trained on the same near-tie shadow.	Residual-bias students achieve more than 50% of the LoRA students' HellaSwag transfer from the same shadow, each against its own shuffle control. Requires Delta(LoRA) >= 1.0 pp with a 95% interval excluding 0. Otherwise the event does not occur and the run is recorded as inconclusive.	0.17	false	false	false	opus	RT	2609.29233	opus-team	4	1	greater	0.5	composite	30	true	Qwen/Qwen2.5-1.5B-Instruct	1	true	1	sol	sol	a100-large	a100-large	submitted	0	0	1,150	sol	submitted	25.2	22	0.597	0.92	21,383	0.1	null	null	null	null	null	null	null	null	distinct	null	distinct	false	faithful	valid	1	3	unfaithful	valid	2	3	faithful	valid	1	3	true	1	3	false	true	false	5	0.16	4	0.05	5	0.08	5	0.25	4	0.34	5	0.1	4	neither	builds on	some	a few groups	no	only if	failed	inconclusive	repair	null	neither	explains	few	a few groups	no	only if	failed	inconclusive	repair	null	neither	builds on	some	a few groups	no	only if	failed	inconclusive	repair	null	builds on	false	true	false	false	inconclusive	repair	0
Exact ancestor or shared pretraining? ATD transfer to Qwen2.5-1.5B base from a teacher post-trained on Qwen2.5-1.5B-Instruct, with a base-plus-update reference	A student initialized from Qwen2.5-1.5B base and trained on the same ATD labels keeps more than half of the signal-minus-shuffle HellaSwag transfer of a student initialized from the exact instruct ancestor.	The pretrained-base student keeps more than half of the HellaSwag transfer obtained from the exact post-trained ancestor, each against a shuffle control with the same initialization. Requires Delta_M0 >= 1.0 pp with a 95% interval excluding 0. Otherwise the event does not occur and the run is recorded as inconclusive.	0.3	true	true	false	opus	RT	2609.29233	opus-team	5	1	greater	0.5	composite	32	true	Qwen/Qwen2.5-1.5B-Instruct; Qwen/Qwen2.5-1.5B	1	true	1	sol	sol	a100-large	a100-large	submitted	0.895238	0.895238	1,324	sol	submitted	27.7	27	0.633	1.046	20,799	0.1	null	null	null	null	null	null	null	null	distinct	null	distinct	true	faithful	valid	2	2	faithful	valid	2	2	faithful	valid	2	1	true	2	1.67	true	true	true	3	0.3	5	0.07	1	0.15	2	0.31	5	0.38	2	0.2	3	neither	builds on	some	a few groups	no	only if	held	inconclusive	repair	null	neither	checks	some	one topic	no	only if	held	for	one more experiment	null	neither	checks	some	a few groups	no	only if	held	for	larger study	null	checks	true	false	false	true	for	null	0
Update-source identification from discarded-response masks	The pattern of admissible versus inadmissible responses identifies a teacher's post-training task on independently trained checkpoints, beyond differences in overall retention rates.	Retention-mask source identification exceeds the rate- and difficulty-preserving null by more than ten percentage points.	0.42	false	false	true	sol	SR	2609.29233	sol-single	1	1	greater	0.1	confidence bound	26	true	Qwen/Qwen2.5-1.5B-Instruct with ten separately trained rank-8 LoRA adapters	1	true	1	opus	opus	a100-large	a100-large	submitted	0	0	411	opus	submitted	10.3	11	1.427	0.346	18,479	0.1	null	null	null	null	null	null	null	null	distinct	false	distinct	false	faithful	valid	1	5	unfaithful	valid	1	5	faithful	valid	1	5	true	1	5	false	true	false	1	0.7	1	0.35	3	0.15	4	0.62	1	0.1	5	0.15	5	neither	builds on	few	a few groups	no	only if	failed	inconclusive	repair	null	neither	builds on	few	a few groups	no	only if	failed	inconclusive	repair	null	neither	builds on	few	one topic	no	only if	held	inconclusive	repair	null	builds on	false	true	true	false	inconclusive	repair	0
Precision-only changes as a null behavioral shadow	Changing only inference precision produces at least one quarter as many next-token disagreements as the released coding teacher's update on the same eligible carrier population.	The precision-only disagreement burden is demonstrably at least one quarter of the released post-training disagreement burden.	0.76	false	false	true	sol	SR	2609.29233	sol-single	2	1	greater	0.25	confidence bound	16	true	Qwen/Qwen2.5-1.5B-Instruct evaluated in BF16 and FP32	1	true	1	opus	opus	a100-large	a100-large	submitted	0	0	125	opus	submitted	6.9	13	1.474	0.238	16,246	0.1	null	null	null	null	null	null	null	null	distinct	false	distinct	false	faithful	valid	1	4	faithful	valid	2	4	faithful	valid	1	4	true	1	4	false	true	false	4	0.83	3	0.62	1	0.75	2	0.78	2	0.4	2	0.22	3	neither	builds on	few	a few groups	no	only if	failed	inconclusive	repair	null	neither	builds on	few	a few groups	no	only if	failed	inconclusive	repair	null	neither	builds on	some	a few groups	no	only if	held	for	larger study	null	builds on	false	true	true	false	inconclusive	repair	0
Does coding transfer survive a generic system prompt?	Replacing the named Qwen system prompt with a generic helpful-assistant prompt preserves at least half of ATD's coding advantage over its exact matched control.	A positive ATD coding advantage is present under the named prompt, and the generic prompt retains more than half of that advantage.	0.48	false	false	false	sol	SR	2609.29233	sol-single	3	1	greater	0	confidence bound	29	true	Qwen/Qwen2.5-1.5B-Instruct with separately trained signal and exact-control rank-16 LoRA adapters	1	true	1	opus	opus	a100-large	a100-large	submitted	-0.018293	-0.018293	862	opus	submitted	33.8	19	1.81	1.33	23,778	0.1	null	null	null	null	null	null	null	null	distinct	false	distinct	false	faithful	valid	1	3	faithful	valid	1	3	faithful	valid	2	3	true	1	3	false	true	false	5	0.32	4	0.12	2	0.08	3	0.53	4	0.48	4	0.4	2	neither	checks	some	a few groups	no	only if	held	inconclusive	repair	null	neither	varies	few	a few groups	no	only if	held	inconclusive	repair	null	neither	builds on	some	a few groups	no	only if	held	inconclusive	repair	null	null	false	false	false	false	inconclusive	repair	0
Transplanting a coding shadow as one activation vector	A single mean residual-stream shift measured on unrelated prompts is sufficient to recover at least half of the ATD coding advantage when transplanted into the exact-control student.	Baseline ATD transfer is positive, the transplanted mean vector recovers more than half its benefit, and it outperforms the norm-matched random intervention.	0.14	false	true	false	sol	SR	2609.29233	sol-single	4	1	greater	0	confidence bound	31	true	Qwen/Qwen2.5-1.5B-Instruct with signal and exact-control rank-16 LoRA adapters, including two recipient activation interventions	1	true	1	opus	opus	a100-large	a100-large	submitted	-0.198171	-0.198171	814	opus	submitted	26.7	24	1.957	1.019	24,228	0.1	null	null	null	null	null	null	null	null	distinct	false	distinct	false	faithful	valid	2	1	unfaithful	valid	2	1	faithful	valid	2	1	true	2	1	true	true	false	2	0.07	5	0.04	5	0.02	5	0.1	5	0.08	3	0.08	4	neither	builds on	some	one topic	no	only if	held	against	stop	null	neither	explains	some	a few groups	no	only if	held	against	one more experiment	null	neither	explains	some	one topic	no	only if	held	against	one more experiment	null	explains	false	false	false	true	against	one more experiment	1
Selective coding impairment from inverted shadow choices	Complementing the teacher's off-task choices selectively impairs coding relative to a complemented exact control, with a larger impairment on coding than on a held-out science-question endpoint.	The inverted signal causes at least one percentage point of coding impairment beyond its matched control, and that impairment exceeds the science-endpoint impairment by at least one percentage point.	0.31	false	true	false	sol	SR	2609.29233	sol-single	5	1	greater	0.01	confidence bound	27	true	Qwen/Qwen2.5-1.5B-Instruct with complemented-signal and complemented-exact rank-16 LoRA adapters	1	true	1	opus	opus	a100-large	a100-large	submitted	-0.035488	-0.035488	561	opus	submitted	27.8	30	2.473	1.056	30,253	0.1	null	null	null	null	null	null	null	null	distinct	false	distinct	false	faithful	valid	2	2	faithful	valid	2	2	faithful	valid	1	2	true	2	2	true	true	false	3	0.24	2	0.08	4	0.09	1	0.24	3	0.52	1	0.35	1	neither	builds on	some	a few groups	no	yes	held	inconclusive	repair	null	neither	builds on	few	a few groups	no	only if	cannot tell	inconclusive	repair	null	neither	builds on	some	a few groups	no	only if	held	inconclusive	repair	null	builds on	false	true	false	false	inconclusive	repair	0
Near-ties versus modest margins for observing the strength of a fixed update	At equal one-token query budgets, modest-margin probes distinguish half-strength from full-strength copies of one fixed update more often than near-tie probes.	Modest-margin probes yield at least three percentage points more usable strength distinctions per equal-cost query panel, with a positive uncertainty margin.	0.25	false	true	true	sol	RT	2609.29233	sol-team	1	1	greater	0.03	confidence bound	29	true	Qwen/Qwen2.5-1.5B-Instruct with one rank-8 OpenBookQA adapter evaluated at strengths 0, 0.5 and 1	1	true	1	opus	opus	a100-large	a100-large	submitted	-0.043902	-0.043902	166	opus	submitted	8.6	11	1.461	0.28	21,076	0.1	null	null	null	null	null	null	null	null	distinct	null	distinct	true	faithful	valid	2	4	faithful	valid	2	5	faithful	valid	2	4	true	2	4.33	true	true	true	4	0.33	4	0.2	3	0.29	2	0.33	3	0.18	2	0.22	3	neither	builds on	few	a few groups	no	yes	held	inconclusive	repair	null	neither	builds on	few	a few groups	no	yes	held	inconclusive	repair	null	neither	builds on	few	a few groups	no	only if	held	inconclusive	repair	null	builds on	false	true	true	false	inconclusive	repair	0
Identifying a newly memorized random bit through unrelated hard-token signatures	Unrelated greedy word-choice signatures identify which of two balanced random factual mappings a teacher learned, on held-out optimization seeds.	After adequate unrestricted recall learning, unrelated signatures identify the learned key above 75% held-out balanced accuracy and above the paired teacher-level permutation null.	0.35	true	true	false	sol	RT	2609.29233	sol-team	2	1	greater	0	composite	29	true	Qwen/Qwen2.5-0.5B-Instruct; Twelve short rank-16 LoRA teachers with synthetic factual bindings	1	true	1	opus	opus	a100-large	a100-large	submitted	0.166667	0.166667	316	opus	submitted	8.1	8	1.116	0.282	16,036	0.1	null	null	null	null	null	null	null	null	distinct	null	distinct	false	faithful	valid	2	5	faithful	valid	2	3	unknown	unknown	null	5	true	2	4.33	true	true	false	1	0.3	1	0.2	2	0.06	3	0.42	1	0.09	4	0.07	1	neither	builds on	some	one topic	no	only if	held	for	larger study	null	neither	builds on	few	a few groups	no	only if	held	for	larger study	null	neither	builds on	some	one topic	no	only if	held	for	larger study	null	builds on	false	true	false	true	for	larger study	0
A pair-aware predictive audit of the released coding shadow	Public context features reduce source-held-out teacher-choice log loss beyond an additive lexical ranking, an ordered-pair-aware static predictor, and a matched-capacity same-pair context-shuffled predictor.	Contextual prediction has a nontrivial source-held-out advantage over all three registered controls.	0.4	false	true	true	sol	RT	2609.29233	sol-team	3	1	greater	0.01	confidence bound	23	true	Qwen/Qwen2.5-1.5B-Instruct at the pinned released revision	1	true	1	opus	opus	a100-large	a100-large	submitted	-0.044419	-0.044419	46	opus	submitted	7.8	15	1.806	0.237	24,596	0.1	null	null	null	null	null	null	null	null	distinct	null	distinct	false	faithful	valid	2	3	faithful	valid	2	4	faithful	valid	2	3	true	2	3.33	true	true	false	2	0.47	2	0.15	4	0.04	4	0.36	2	0.32	3	0.27	4	neither	builds on	some	a few groups	no	yes	held	against	stop	null	neither	builds on	few	a few groups	no	only if	cannot tell	against	one more experiment	null	neither	builds on	few	a few groups	no	yes	held	against	one more experiment	null	builds on	false	true	true	true	against	one more experiment	0
A controlled numerical-resolution audit of released private-update bits	More than 20% of released teacher-versus-public disagreement rows have opposite unrestricted public word choices under FP32 and BF16 backbone arithmetic at the same batch size, despite a common FP32 output head.	A substantial fraction of recorded private-update disagreements changes binary interpretation under precision alone at fixed batching and fixed FP32 readout.	0.2	true	true	false	sol	RT	2609.29233	sol-team	4	1	greater	0.2	confidence bound	27	true	Qwen/Qwen2.5-1.5B-Instruct executed with FP32 and BF16 backbone arithmetic	1	true	1	opus	opus	a100-large	a100-large	submitted	0.408665	0.408665	356	opus	submitted	9.7	11	1.537	0.331	18,641	0.1	null	null	null	null	null	null	null	null	distinct	null	distinct	false	faithful	valid	2	1	faithful	valid	2	1	faithful	valid	3	1	true	2	1	true	true	false	5	0.51	3	0.42	1	0.38	1	0.39	4	0.35	1	0.55	2	neither	builds on	few	a few groups	no	yes	held	for	stop	null	neither	checks	few	a few groups	no	yes	held	for	one more experiment	null	neither	builds on	some	a few groups	no	yes	held	for	larger study	null	builds on	false	true	true	true	for	null	1
Boundary-targeted public-base anchoring for reducing a teacher's behavioral shadow	Boundary-targeted anchoring achieves a better held-out shadow-versus-science-utility point than equal-budget passive anchoring.	With both teachers showing adequate utility and shadow changes, active anchoring has a jointly resolved reduction versus sham and passive anchoring while preserving the registered science-utility margins.	0.1	false	true	false	sol	RT	2609.29233	sol-team	5	1	greater	0	confidence bound	30	true	Qwen/Qwen2.5-0.5B-Instruct; Two science LoRA teachers and three continuation arms per teacher	1	true	1	opus	opus	a100-large	a100-large	submitted	-0.25	-0.25	205	opus	submitted	7.3	9	1.358	0.205	23,574	0.1	null	null	null	null	null	null	null	null	distinct	null	distinct	false	faithful	valid	2	2	faithful	valid	2	2	faithful	valid	2	2	true	2	2	true	true	false	3	0.06	5	0.05	5	0.03	5	0.09	5	0.08	5	0.1	5	neither	builds on	some	a few groups	no	only if	held	against	stop	null	neither	builds on	few	a few groups	no	only if	held	against	stop	null	neither	builds on	some	one topic	no	only if	held	inconclusive	repair	null	builds on	false	true	false	true	against	stop	0
Within-band margin value for one-token ATD	At equal retained-row count and training compute, the tightest-margin carriers yield a larger signal-minus-shuffle HumanEval+ gain than the loosest-margin carriers.	The event occurs if the tight-margin group has a larger signal-minus-shuffle HumanEval+ gain than the loose-margin group.	0.62	true	false	false	luna	SR	2609.29233	luna-single	1	1	greater	0	plain threshold	28	true	Qwen/Qwen2.5-1.5B-Instruct	1	true	1	sol	sol	a100-large	a100-large	submitted	1.829268	1.829268	344	sol	submitted	24.2	33	0.765	0.922	18,420	0.1	null	null	null	null	null	null	null	null	distinct	null	distinct	true	faithful	valid	1	4	faithful	valid	1	4	faithful	valid	1	3	true	1	3.67	false	true	true	2	0.48	5	0.42	5	0.44	5	0.53	5	0.66	1	0.62	1	neither	builds on	some	a few groups	no	yes	held	inconclusive	repair	null	neither	builds on	some	a few groups	no	only if	cannot tell	inconclusive	repair	null	neither	builds on	most	a few groups	no	yes	held	inconclusive	repair	null	builds on	false	true	false	false	inconclusive	repair	0
Unique-carrier data scale at fixed student compute	At a fixed student update budget, training on all 5,664 unique released carriers yields a larger signal-minus-shuffle HumanEval+ gain than training on 1,024 unique carriers.	The event occurs if the signal-minus-shuffle HumanEval+ gain for 5,664 rows exceeds the gain for 1,024 rows.	0.67	true	false	false	luna	SR	2609.29233	luna-single	2	1	greater	0	plain threshold	30	true	Qwen/Qwen2.5-1.5B-Instruct	1	true	1	sol	sol	a100-large	a100-large	submitted	2.439024	2.439024	469	sol	submitted	33	37	0.813	1.257	22,943	0.1	null	null	null	null	null	null	null	null	distinct	null	close variant	true	faithful	valid	1	3	faithful	valid	1	3	faithful	valid	1	4	true	1	3.33	false	false	true	3	0.61	3	0.49	2	0.47	4	0.69	2	0.61	2	0.68	4	neither	builds on	some	a few groups	no	yes	held	inconclusive	repair	null	neither	builds on	most	a few groups	no	only if	held	inconclusive	repair	null	neither	builds on	most	a few groups	no	yes	held	inconclusive	repair	null	builds on	false	true	false	false	inconclusive	repair	0
Student-side robustness to a carrier-instruction paraphrase	Re-expressing the carrier instructions in a meaning-preserving template while retaining the original teacher choices still yields a positive signal-minus-exact-control HumanEval+ gain.	The event occurs if the paraphrased-prompt signal adapter outperforms its exact nuisance-matched control on HumanEval+ pass@1.	0.42	true	false	false	luna	SR	2609.29233	luna-single	3	1	greater	0	plain threshold	24	true	Qwen/Qwen2.5-1.5B-Instruct	1	true	1	sol	sol	a100-large	a100-large	submitted	2.439024	2.439024	574	sol	submitted	25	28	0.581	0.975	14,736	0.1	null	null	null	null	null	null	null	null	distinct	null	distinct	false	faithful	valid	1	2	faithful	valid	1	2	faithful	valid	1	2	true	1	2	false	true	false	4	0.68	4	0.54	3	0.53	1	0.74	4	0.72	4	0.75	5	neither	builds on	some	a few groups	no	yes	held	inconclusive	repair	null	neither	varies	some	a few groups	no	only if	held	inconclusive	repair	null	neither	builds on	some	a few groups	no	yes	held	inconclusive	repair	null	builds on	false	true	false	false	inconclusive	repair	0
Capability transfer to a smaller sibling student	The released code-shadow prompt-label pairing yields a positive HumanEval+ gain over its matched control when the student is Qwen2.5-0.5B-Instruct rather than the Qwen2.5-1.5B ancestor.	The event occurs if the smaller sibling student's signal adapter outperforms its exact nuisance-matched control on HumanEval+ pass@1.	0.45	false	true	false	luna	SR	2609.29233	luna-single	4	1	greater	0	plain threshold	20	true	Qwen/Qwen2.5-0.5B-Instruct	1	true	1	sol	sol	a100-large	a100-large	submitted	0	0	383	sol	submitted	42	30	0.614	1.675	17,578	0.1	null	null	null	null	null	null	null	null	distinct	null	distinct	true	faithful	valid	1	1	faithful	valid	2	1	faithful	valid	2	1	true	2	1	true	true	true	1	0.42	1	0.38	1	0.42	3	0.37	1	0.53	3	0.44	2	neither	builds on	most	a few groups	no	yes	held	inconclusive	repair	null	neither	builds on	some	one topic	no	only if	cannot tell	inconclusive	repair	null	neither	builds on	most	a few groups	no	yes	held	inconclusive	repair	null	builds on	false	true	false	false	inconclusive	repair	1
Does ATD's pairing-specific gain depend on AdamW?	After matching the first-step global adapter update norm, AdamW produces a larger pairing-specific HumanEval+ gain than SGD with momentum on the ATD carriers.	The event occurs if AdamW's signal-minus-exact-control HumanEval+ gain exceeds SGD's signal-minus-exact-control gain.	0.66	false	false	false	luna	SR	2609.29233	luna-single	5	1	greater	0	plain threshold	32	true	Qwen/Qwen2.5-1.5B-Instruct	1	true	1	sol	sol	a100-large	a100-large	submitted	-0.609756	-0.609756	597	sol	submitted	17.6	22	0.494	0.638	14,791	0.1	null	null	null	null	null	null	null	null	distinct	null	distinct	false	faithful	valid	1	5	faithful	valid	1	5	faithful	valid	1	5	true	1	5	false	true	false	5	0.64	2	0.46	4	0.53	2	0.67	3	0.59	5	0.71	3	neither	builds on	some	a few groups	no	yes	held	inconclusive	repair	null	neither	builds on	few	a few groups	no	only if	failed	inconclusive	repair	null	neither	explains	some	a few groups	no	yes	held	inconclusive	repair	null	builds on	false	true	false	false	inconclusive	repair	0
Unique ATD coverage versus replay at a fixed update budget	At the same student update budget, training on all 5,664 rows will produce a larger exact-control-corrected HumanEval+ gain than training on a nested 2,048-row subset.	The pilot supports greater transfer from broader unique-row coverage if the lower endpoint of the paired task-bootstrap interval for the full-data transfer gain minus the subset transfer gain is above zero.	0.34	false	false	false	luna	RT	2609.29233	luna-team	1	1	greater	0	confidence bound	32	true	Qwen/Qwen2.5-1.5B-Instruct	1	true	1	sol	sol	a100-large	a100-large	submitted	-0.609756	-0.609756	955	sol	submitted	36	33	0.705	1.431	19,372	0.1	null	null	null	null	null	null	null	null	distinct	null	close variant	true	faithful	valid	1	2	faithful	valid	2	1	faithful	valid	2	2	true	2	1.67	true	false	true	2	0.2	2	0.05	3	0.05	4	0.15	4	0.25	2	0.2	4	neither	builds on	few	a few groups	no	yes	held	inconclusive	repair	null	neither	builds on	some	a few groups	no	yes	held	inconclusive	repair	null	neither	builds on	some	a few groups	no	only if	held	inconclusive	repair	null	builds on	false	true	false	false	inconclusive	repair	0
Do tighter public near-ties carry more coding-transfer signal?	With equal row counts and student-update budgets, the tightest-margin quartile will yield a larger exact-control-corrected HumanEval+ gain than the loosest-margin quartile.	The pilot supports greater transfer from tighter near-ties if the lower endpoint of the paired task-bootstrap interval for tight-group gain minus loose-group gain is above zero.	0.22	false	true	false	luna	RT	2609.29233	luna-team	2	1	greater	0	confidence bound	32	true	Qwen/Qwen2.5-1.5B-Instruct	1	true	1	sol	sol	a100-large	a100-large	submitted	-4.878049	-4.878049	1,155	sol	submitted	36.7	36	0.772	1.444	21,501	0.1	null	null	null	null	null	null	null	null	distinct	null	distinct	true	faithful	valid	1	3	faithful	valid	2	4	faithful	valid	2	4	true	2	3.67	true	true	true	4	0.08	5	0.04	5	0.03	5	0.07	5	0.22	4	0.12	2	neither	builds on	some	a few groups	no	only if	held	inconclusive	repair	null	neither	builds on	some	a few groups	no	only if	held	inconclusive	repair	null	neither	builds on	most	a few groups	no	only if	held	inconclusive	repair	null	builds on	false	true	false	false	inconclusive	repair	0
Can the first transformer block alone recover a coding shadow?	A student trained with LoRA restricted to the first transformer block will achieve a positive exact-control-corrected HumanEval+ gain.	The pilot supports a first-block coding shadow if the lower endpoint of the paired task-bootstrap interval for signal minus exact control is above zero.	0.2	false	true	false	luna	RT	2609.29233	luna-team	3	1	greater	0	confidence bound	28	true	Qwen/Qwen2.5-1.5B-Instruct	1	true	1	sol	sol	a100-large	a100-large	submitted	-6.097561	-6.097561	605	sol	submitted	19.2	23	0.427	0.763	9,754	0.1	null	null	null	null	null	null	null	null	distinct	null	distinct	false	faithful	valid	1	4	faithful	valid	2	3	faithful	valid	2	3	true	2	3.33	true	true	false	5	0.14	4	0.08	2	0.1	3	0.14	2	0.1	5	0.15	5	neither	builds on	some	a few groups	no	yes	held	inconclusive	repair	null	neither	builds on	few	a few groups	no	only if	held	against	one more experiment	null	neither	builds on	some	a few groups	no	yes	held	inconclusive	repair	null	builds on	false	true	false	false	inconclusive	repair	0
Does the learned carrier-choice signal generalize to unseen contexts under a reworded instruction?	On carrier contexts withheld from student training, the signal student's margin toward the original teacher label will exceed the exact-control student's margin when the choice instruction is reworded.	The pilot supports persistence under the rewrite if the lower endpoint of the source-group bootstrap interval for the paraphrased signal-minus-control margin is above zero, provided the original-format held-out margin check also has a lower endpoint above zero.	0.4	true	true	false	luna	RT	2609.29233	luna-team	4	1	greater	0	confidence bound	25	true	Qwen/Qwen2.5-1.5B-Instruct	1	true	1	sol	sol	a100-large	a100-large	submitted	0.139629	0.139629	434	sol	submitted	13.1	11	0.361	0.496	8,223	0.1	null	null	null	null	null	null	null	null	distinct	null	distinct	false	faithful	valid	2	1	unfaithful	valid	2	2	faithful	valid	2	1	true	2	1.33	true	true	false	1	0.7	1	0.35	1	0.62	1	0.45	1	0.36	1	0.38	1	neither	builds on	some	a few groups	no	yes	held	for	larger study	null	neither	builds on	few	a few groups	no	only if	held	for	one more experiment	null	neither	builds on	some	a few groups	no	only if	held	for	one more experiment	null	builds on	false	true	false	true	for	one more experiment	1
Can a safety-SFT update transmit a safety preference through ATD?	Provided a locally trained safety teacher improves on held-out safety preferences, a student trained only on its near-tie carrier choices will rank held-out safe responses above unsafe responses more often than an exact-matched control.	The pilot supports safety-preference transfer if at least 256 carrier labels are retained, the teacher-validity check passes, and the lower endpoint of the paired prompt-bootstrap interval for signal-minus-control safe-ranking accuracy is above zero.	0.1	false	false	false	luna	RT	2609.29233	luna-team	5	1	greater	0	confidence bound	34	true	Qwen/Qwen2.5-0.5B-Instruct	1	true	1	sol	sol	a100-large	a100-large	submitted	-100	-100	133	sol	submitted	10.3	18	0.493	0.328	15,699	0.1	null	null	null	null	null	null	null	null	distinct	null	distinct	false	faithful	valid	1	5	faithful	invalid	0	5	faithful	valid	1	5	true	1	5	false	true	false	3	0.06	3	0.06	4	0.09	2	0.1	3	0.15	3	0.22	3	neither	builds on	some	one topic	no	only if	failed	inconclusive	repair	null	neither	builds on	some	one topic	no	only if	failed	inconclusive	repair	null	neither	varies	most	one topic	no	only if	failed	inconclusive	repair	null	builds on	false	true	false	false	inconclusive	repair	0
Near-tie enrichment of base-instruct disagreement on unrelated forced choices	Near-tie carriers show higher greedy teacher-ancestor disagreement than far carriers.	Near-tie flip rate exceeds far flip rate by more than 5 percentage points	0.7	true	false	false	muse	SR	2609.29233	muse-single	1	1	greater	0.05	plain threshold	22	true	Qwen/Qwen2.5-0.5B; Qwen/Qwen2.5-0.5B-Instruct	1	true	1	opus	opus	a100-large	a100-large	submitted	0.192344	0.192344	75	opus	submitted	3.5	11	1.086	0.141	9,656	0.1	null	null	null	null	null	null	null	null	distinct	null	distinct	false	faithful	valid	1	3	unfaithful	valid	1	1	unfaithful	valid	1	3	false	1	2.33	false	true	false	2	0.7	1	0.75	3	0.7	3	0.74	3	0.72	2	0.85	1	neither	checks	most	a few groups	no	yes	held	for	larger study	null	neither	builds on	some	a few groups	no	yes	held	for	one more experiment	null	neither	checks	most	a few groups	no	yes	held	for	one more experiment	null	checks	true	false	false	true	for	one more experiment	0
Single-sample temperature robustness of greedy behavioral shadows	A single temperature-1 sample from the teacher agrees with its greedy choice on near-tie carriers at high rate.	Single temperature-1 sample matches greedy teacher token on more than 70 percent of retained near-tie carriers	0.3	true	true	false	muse	SR	2609.29233	muse-single	2	1	greater	0.7	plain threshold	20	true	Qwen/Qwen2.5-0.5B; Qwen/Qwen2.5-0.5B-Instruct	1	true	1	opus	opus	a100-large	a100-large	submitted	0.763052	0.763052	36	opus	submitted	4.1	12	1.131	0.149	9,666	0.1	null	null	null	null	null	null	null	null	distinct	null	distinct	false	faithful	valid	1	2	unfaithful	valid	2	4	faithful	valid	2	1	true	2	2.33	true	true	false	4	0.1	5	0.47	2	0.48	2	0.08	5	0.22	4	0.2	4	neither	builds on	some	a few groups	no	yes	held	for	one more experiment	null	neither	builds on	some	a few groups	no	yes	held	for	one more experiment	null	neither	builds on	some	a few groups	no	yes	held	for	one more experiment	null	builds on	false	true	false	true	for	one more experiment	0
Task fingerprinting from one-bit unrelated word choices without distillation	Per-word choice rates on shared near-tie carriers identify which task a same-ancestor teacher was trained on above chance.	Held-out teacher identification accuracy from per-word rates exceeds 60 percent	0.25	false	false	false	muse	SR	2609.29233	muse-single	3	1	greater	0.6	plain threshold	30	true	Qwen/Qwen2.5-0.5B-Instruct	1	true	1	opus	opus	a100-large	a100-large	submitted	0.506024	0.506024	67	opus	submitted	3.4	10	1.099	0.113	10,978	0.1	null	null	null	null	null	null	null	null	distinct	null	distinct	false	faithful	valid	1	5	unfaithful	valid	1	5	unfaithful	valid	1	5	false	1	5	false	true	false	5	0.18	4	0.28	1	0.03	4	0.24	4	0.08	5	0.15	5	neither	builds on	some	a few groups	no	only if	failed	inconclusive	repair	null	neither	builds on	some	a few groups	no	only if	failed	inconclusive	repair	null	neither	builds on	some	a few groups	no	yes	held	inconclusive	repair	null	builds on	false	true	false	false	inconclusive	repair	0
Paraphrase robustness of near-tie flips	Flips versus base on the original template predict flips on a paraphrased template above the paraphrase base rate.	Flip preservation lift over paraphrase base rate exceeds 15 percentage points	0.55	false	false	false	muse	SR	2609.29233	muse-single	4	1	greater	0.15	plain threshold	22	true	Qwen/Qwen2.5-0.5B; Qwen/Qwen2.5-0.5B-Instruct	1	true	1	opus	opus	a100-large	a100-large	submitted	0.050861	0.050861	72	opus	submitted	3.2	12	1.145	0.126	9,423	0.1	null	null	null	null	null	null	null	null	distinct	null	distinct	false	faithful	valid	1	1	faithful	valid	1	2	faithful	valid	2	2	true	1	1.67	false	true	false	3	0.2	3	0.13	5	0.05	5	0.36	2	0.55	3	0.6	2	neither	builds on	some	a few groups	no	yes	held	inconclusive	repair	null	neither	builds on	some	a few groups	no	only if	failed	inconclusive	repair	null	neither	builds on	some	a few groups	no	only if	failed	inconclusive	repair	null	builds on	false	true	false	false	inconclusive	repair	1
Tiny-budget taskless transfer on multiple-choice science with likelihood eval	Signal student exceeds shuffle student on held-out OpenBookQA accuracy.	Signal minus shuffle accuracy on held-out OpenBookQA exceeds 1 percentage point	0.38	false	false	false	muse	SR	2609.29233	muse-single	5	1	greater	1	plain threshold	30	true	Qwen/Qwen2.5-0.5B-Instruct	1	true	1	opus	opus	a100-large	a100-large	submitted	-0.2	-0.2	80	opus	submitted	4.8	11	1.181	0.16	11,505	0.1	null	null	null	null	null	null	null	null	close variant	null	close variant	false	faithful	valid	1	4	faithful	valid	1	3	faithful	valid	1	4	true	1	3.67	false	false	false	1	0.23	2	0.17	4	0.17	1	0.2	1	0.34	1	0.35	3	neither	builds on	some	a few groups	no	only if	held	inconclusive	repair	null	neither	varies	some	a few groups	no	only if	held	inconclusive	repair	null	reported	varies	some	a few groups	no	only if	held	inconclusive	repair	true	varies	false	false	false	false	inconclusive	repair	0
Tight near-ties flip more than far pairs under a small task delta	Greedy flip rate versus the public ancestor is higher on tight near-ties than on far-from-tie pairs for the same small SFT delta.	Greedy flip rate on tight near-ties exceeds flip rate on far pairs by more than 5pp	0.8	true	false	false	muse	RT	2609.29233	muse-team	1	1	greater	0.05	plain threshold	25	true	Qwen/Qwen2.5-0.5B-Instruct	1	true	1	opus	opus	a100-large	a100-large	submitted	0.43429	0.43429	126	opus	submitted	5.8	6	0.867	0.233	9,486	0.1	null	null	null	null	null	null	null	null	distinct	null	distinct	true	faithful	valid	1	4	faithful	valid	1	4	faithful	valid	2	4	true	1	4	false	true	true	2	0.86	1	0.85	3	0.91	4	0.82	1	0.78	1	0.85	2	neither	checks	some	a few groups	no	only if	held	for	stop	null	neither	checks	some	a few groups	no	only if	held	for	stop	null	neither	checks	most	a few groups	no	yes	held	for	one more experiment	null	checks	true	false	false	true	for	stop	0
Task LoRA shadow transfers more than norm-matched random LoRA	Task-trained signal-minus-shuffle exceeds random-perturbation signal-minus-shuffle on held-out SciQ.	Task-LoRA signal-minus-shuffle exceeds random-LoRA signal-minus-shuffle by more than 1pp	0.55	true	true	false	muse	RT	2609.29233	muse-team	2	1	greater	0.01	plain threshold	32	true	Qwen/Qwen2.5-0.5B-Instruct	1	true	1	opus	opus	a100-large	a100-large	submitted	0.076	0.076	177	opus	submitted	8.7	5	1.104	0.302	18,517	0.1	null	null	null	null	null	null	null	null	distinct	null	distinct	true	faithful	valid	2	1	unfaithful	valid	2	1	faithful	valid	2	1	true	2	1	true	true	true	1	0.32	5	0.38	2	0.26	3	0.38	5	0.51	2	0.5	1	neither	checks	some	one topic	no	only if	held	for	larger study	null	neither	checks	some	a few groups	no	only if	held	for	larger study	null	neither	checks	some	a few groups	no	no	held	for	one more experiment	null	checks	true	false	false	true	for	larger study	1
Same-task shadows agree more than different-task shadows	Two math teachers trained on disjoint data agree with each other more than either agrees with a science teacher on shared near-ties.	Within-task agreement exceeds mean between-task agreement by more than 2pp	0.6	true	false	false	muse	RT	2609.29233	muse-team	3	1	greater	0.02	plain threshold	28	true	Qwen/Qwen2.5-0.5B-Instruct	1	true	1	opus	opus	a100-large	a100-large	submitted	0.270634	0.276281	185	opus	submitted	10.8	7	0.954	0.423	12,237	0.1	null	null	null	null	null	null	null	null	distinct	null	distinct	true	faithful	valid	2	3	unfaithful	valid	2	3	unfaithful	valid	1	3	false	2	3	true	true	true	3	0.8	2	0.72	1	0.86	2	0.74	2	0.53	3	0.65	3	neither	builds on	some	a few groups	no	yes	held	for	one more experiment	null	neither	builds on	few	a few groups	no	yes	held	for	one more experiment	null	neither	builds on	some	a few groups	no	yes	held	for	one more experiment	null	builds on	false	true	false	true	for	one more experiment	0
Near-tie queries do not port across model scale	Most near-ties selected by one scale are far from tie under the other scale.	Fraction of 0.5B near-ties that remain near-ties under 1.5B is below 30 percent	0.85	true	false	true	muse	RT	2609.29233	muse-team	4	1	less	0.3	plain threshold	20	true	Qwen/Qwen2.5-0.5B-Instruct; Qwen/Qwen2.5-1.5B-Instruct	1	true	1	opus	opus	a100-large	a100-large	submitted	0.04499	0.04499	98	opus	submitted	5.1	6	0.902	0.214	10,556	0.1	null	null	null	null	null	null	null	null	distinct	null	distinct	false	faithful	valid	2	2	unfaithful	valid	2	2	unfaithful	valid	2	2	false	2	2	true	true	false	4	0.96	3	0.88	5	0.96	5	0.93	3	0.72	5	0.9	5	neither	builds on	few	a few groups	no	yes	held	for	stop	null	neither	builds on	few	a few groups	no	only if	held	for	stop	null	neither	builds on	some	a few groups	no	yes	held	for	larger study	null	builds on	false	true	true	true	for	stop	0

End of preview. Expand in Data Studio

rTaste

rTaste measures the early stage research judgment of six frontier models. Given a recent NLP paper, each model, working alone or as a team, proposes five minimum viable experiments (MVEs), forecasts whether each will pass a threshold it sets and ranks them. A separate executor then attempts every MVE, including those ranked last, and model auditors judge the results. The dataset holds all 415 MVEs based on seven seed papers, with their forecasts, rankings and audited outcomes.

Paper: rTaste: Do Research Agents Have Taste? Evaluating Early Stage Research Judgment
Code: github.com/ArtKanke/rTaste

Configs

Config	Rows	Data
`mves` (default)	415	one MVE, with its hypothesis, every label and every forecast
`mves_full`	415	one MVE, with the full text of its registration and of every judgment
`seeds`	7	one seed paper
`conditions`	84	one model and harness on a seed paper
`rankings`	83	one condition's rankings of its five MVEs, with the reasons
`follow_up_lists`	42	the ten follow ups one model listed for one seed paper
`calls`	20,593	one model call or machine
`auditor_check`	10	one copy of an MVE with a planted defect

The configs join on seed, condition and mve. A null means there is no value.

How it was built

Propose. On each seed paper, six models register five MVEs each, once as a single researcher and once as a research team.
Forecast. Each agent gives every MVE a probability that its registered event will occur and ranks the five. Six judges do the same from the seed paper and the registrations alone. Two novelty judges (Claude Opus 5.5, GPT-6.1 Sol) compare each MVE with prior work and with lists of obvious follow ups.
Run. One of two executors (Claude Opus 5.5, GPT-6.1 Sol) implements each MVE on one GPU, and a rerun on a new machine decides whether the event occurred.
Assess. Three auditors (Claude Opus 5.5, GPT-6.1 Sol, Grok 4.7) judge whether each MVE is faithful and valid and rate its research value. After the study, three reviewers, the same models, judge the idea and the evidence.

Fields

The main columns of mves. hypothesis and event_description are copied from mves_full; the other columns are defined in data/README.md.

Column	Meaning
`title`, `hypothesis`, `event_description`	what the MVE tests and its registered event
`agent_probability`, `agent_rank`	the agent's forecast and rank (1 is the MVE most worth pursuing)
`judge_<model>_probability`, `judge_<model>_rank`	the same from each of the six judges
`event`	the registered event occurred in the rerun
`value`, `relation`, `threshold`	the rerun's statistic, and the registered relation and threshold
`delivered`	a rerun produced the registered statistic
`sound`	at least two of three auditors say faithful and at least two say valid
`research_value`	median of the auditors' research values, 0 to 3
`new`	both novelty judges say distinct
`good`	delivered, sound, valuable (`research_value` of 2 or more) and new
`on_list`	on a list of obvious follow ups
`bno_2of3`	at least two reviewers say beyond the paper and non obvious
`informative_2of3`	at least two reviewers read the evidence as for, against or mixed
`proposer`	`astra` GPT-6 Astra, `fable` Claude Fable 5.1, `opus` Claude Opus 5.5, `sol` GPT-6.1 Sol, `luna` GPT-6 Luna, `muse` Muse Spark 1.3
`harness`	SR (single researcher) or RT (research team)
`seed`, `condition`, `mve`	the seed paper's arXiv id; the model and harness, such as `sol-team`; 1 to 5, the order in which the agent wrote the MVE

Load

from datasets import load_dataset

mves = load_dataset("ArtKanke/rTaste", "mves", split="train", revision="v1.0")
full = load_dataset("ArtKanke/rTaste", "mves_full", split="train", revision="v1.0")

v1.0 is the version behind the paper's numbers.

Example MVE (AURORA-LM seed paper, GPT-6 Astra research team, MVE 1)

Field	Value
Title	Are denoiser residuals unusually harmful after matching decoder-space error geometry?
Hypothesis	Actual two-pass denoiser residuals cause more token-recovery errors than randomized residuals with matched decoder-space magnitudes, clean-latent inner products, and within-block residual correlations.
Registered event	Native residuals increase token error by more than five percentage points relative to geometry-matched randomized residuals.
Agent	probability 0.45, ranked second of five
Same model judge	probability 0.57
Rerun	statistic 0.080 against a threshold of 0.05, so the event occurred
Auditors	sound, research value 2 on a scale of 0 to 3
Novelty judges	new, and on a list of obvious follow ups
Reviewers	beyond the paper and non obvious; the evidence is for the hypothesis; next step: larger study

License

CC BY 4.0. The seed papers and their code belong to their authors; the dataset includes only short quotes from them.

Citation

@misc{kanke2026rtaste,
  title         = {rTaste: Do Research Agents Have Taste? Evaluating Early Stage Research Judgment},
  author        = {Kanke, Art},
  year          = {2026},
  eprint        = {2610.40488},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CL},
  url           = {https://arxiv.org/abs/2610.40488}
}

Downloads last month: -

Paper for ArtKanke/rTaste

rTaste: Do Research Agents Have Taste? Evaluating Early Stage Research Judgment

Paper • 2610.40488 • Published 1 day ago