Evidaxis
Dated record

Trinity-RFT

repo · Post-training & Alignment, n = 16 · agentscope-ai/Trinity-RFT · paper: W4415330145 · as of

0.5 commits/week averaged over the trailing window (niche median 16.05). OpenAlex: 1 citing works; fewer than 3 completed years. Dependents: no reading captured in the partition. 712 GitHub stars, recorded, not scored. Measured for Trinity-RFT as of 2026-11-07; sources: GitHub, OpenAlex and deps.dev where linked.

commits/wk 0.5 averaged, trailing window niche median 16.05 (n = 16) last commit week 2026-10-04 measured · GitHub ·
citations 1 all years, indexed to date niche median 14 (n = 5) measured · OpenAlex ·
direct dependents not captured this week weekly axis input; partition 2026-11-07 niche median 4.5 (n = 6) no reading · deps.dev snapshot ·
stars 712 recorded, not scored niche median 10,496 (n = 16) measured · GitHub ·
Download values (CSV) JSON record Follow this system (feed) Report an error

Is Trinity-RFT still maintained?

0.5 commits per week averaged over the trailing window, under the 5-per-week vote threshold (niche median 16.05/wk).

GitHub averaged 0.5 per week over the trailing window; raw weekly counts below are individual weeks.

Last commit week 2026-10-04; 4 consecutive zero weeks in the captured series.

Commits in the latest captured week vs 4 weeks earlier: -1 · vs 26 weeks earlier: -2 · 52-week range: 0 to 12 commits in an individual week.

line chart: Trinity-RFT, weekly commits, 0 in latest week vs niche median 11.5, 2025-11-09 to 2026-11-01
Trinity-RFT averaged 0.5 commits a week, 0.0 times the niche median of 16.05. GitHub commit_activity · .
Latest eight weeks
Week startingCommitsNiche P25Niche P75
2026-09-1300.7533.25
2026-09-200035.25
2026-09-2741.7537
2026-10-041036.75
2026-10-110033.25
2026-10-1800.7532.25
2026-10-250030
2026-11-0101.534.75
heatmap chart: Trinity-RFT, weeks with commits, 16 of 26, 2026-05-10 to 2026-11-01
16 of the last 26 captured weeks had commits for Trinity-RFT. GitHub commit_activity · .
All 52 weeks
Weekly commits
Week startingCommits
2025-11-098
2025-11-168
2025-11-2312
2025-11-309
2025-12-0710
2025-12-148
2025-12-217
2025-12-285
2026-01-049
2026-01-119
2026-01-185
2026-01-256
2026-02-015
2026-02-089
2026-02-150
2026-02-222
2026-03-012
2026-03-082
2026-03-150
2026-03-222
2026-03-292
2026-04-055
2026-04-120
2026-04-192
2026-04-261
2026-05-032
2026-05-104
2026-05-177
2026-05-2411
2026-05-3110
2026-06-073
2026-06-147
2026-06-217
2026-06-282
2026-07-051
2026-07-121
2026-07-193
2026-07-261
2026-08-020
2026-08-091
2026-08-160
2026-08-230
2026-08-300
2026-09-061
2026-09-130
2026-09-200
2026-09-274
2026-10-041
2026-10-110
2026-10-180
2026-10-250
2026-11-010

GitHub commit_activity · period 2025-11-09 to 2026-11-01 · n = 52 weeks · method weekly totals; reconstructed dates · updated · version 868e3c39d9c7-b1 · CSV

Readings

Recorded values and sources
ReadingValueΔ 4 weeksΔ 26 weeks52-week rangeInstrumentDateReproducePrevious reading
Commits per week0.5-0.4-1.20.5 to 8GitHub, trailing 12 weeks2026-11-07Public query0.6
Citing works1no readingno reading1 to 1OpenAlex, all years2026-11-07Public queryno reading
Weekly direct dependentsno readingno readingno readingno readingWeekly direct dependents2026-11-07Public queryno reading
Daily package dependents (unscored)0no readingno reading0 to 0deps.dev REST, unscored2026-11-06Public query0
Stars7124no reading662 to 712GitHub stars, unscored2026-11-07Public query709

Evidaxis snapshot · period 2026-11-07 · n = 1 system · method source linked readings · updated · version 868e3c39d9c7-b1

How many citations does Trinity-RFT have?

OpenAlex records 1 citing works as of 2026-11-07 (no completed yearly reading); fewer than 3 completed years, so no trend is computed.

OpenAlex citing works by completed calendar year
YearCiting works
Total indexed to date, including current year1

OpenAlex · period 2026-11-07 to 2026-11-07 · n = 0 years · method completed calendar years · updated · version 868e3c39d9c7-b1 · CSV by year

Other indexes count differently. A citation count is the number of citing works an index links to a record on a date; this page counts OpenAlex links to W4415330145 as of 2026-11-07. Check the same work: OpenAlex W4415330145 · Semantic Scholar search · Google Scholar search. Seeing a different number? Report the missing work id.

How to cite Trinity-RFT's record?

Evidaxis. Trinity-RFT measurement record. Snapshot 2026-11-07. urn:evidaxis:claim:e_DQT3XHMXA2C:m5:2026-11-07. CC0-1.0. https://evidaxis.org/e/e_DQT3XHMXA2C/

Evidaxis · period 2026-11-07 · n = 1 system · method claim URN · updated · version 868e3c39d9c7-b1

Is anyone using Trinity-RFT?

Direct dependents: no reading captured in partition not captured; the value is missing, not zero.

Weekly dependents
PartitionDirect dependentsWeekly pointsMethod
no partitionno reading0m5 own packages

deps.dev weekly direct dependents of the system's own packages. A package counts when it is declared in the system's own tree, published under that name and not attributed by deps.dev to another repository. Partition dates can repeat across snapshots. Package linkage is not uniformly verified: packages with no repository metadata (no_link) are unverified and remain admitted.

Daily package dependents (unscored): 0; captured 2026-11-06; package: pypi/trinity-rft. deps.dev REST count for one identity-verified pinned package, captured daily and never scored. Its package scope differs from the weekly package union; neither value replaces the other. 712 GitHub stars, recorded, not scored.

sparkline chart: Trinity-RFT, daily dependents, 0, 2026-w27 to 2026-w45
Trinity-RFT recorded 0 daily dependents (unscored); daily niche median not measured. deps.dev REST (unscored) · .
Captured daily series
Unscored daily dependents, as captured
PeriodDependents
2026-w270
2026-w280
2026-w290
2026-w300
2026-w310
2026-w320
2026-w330
2026-w340
2026-w350
2026-w360
2026-w370
2026-w380
2026-w390
2026-w400
2026-w410
2026-w420
2026-w430
2026-w440
2026-w450

deps.dev · period 2026-11-07 · n = 0 partitions · method m5 own packages · updated · version 868e3c39d9c7-b1 · CSV

How does Trinity-RFT compare with its niche?

16 systems in Post-training & Alignment; niche median 16.05 commits per week. The reader chooses the sort order.

dot plot chart: Trinity-RFT, averaged commits per week, 0.5 vs niche median 16.05, 2026-11-07
Trinity-RFT at 0.5 commits a week among 16 systems; niche median 16.05. GitHub, current cohort · .
All 16 systems, alphabetical; choose a heading to sort
AReaL12no readingno reading5,902
ART20.1no reading110,848
Axolotl20.6no reading212,668
LLaMA-Factory2.5582475,915
ms-swift26.2no reading516,057
NeMo-RL41.5no readingno reading2,117
Open R10.2no readingno reading26,474
Open-Instruct9.7no readingno reading3,924
OpenRLHF2.511no reading10,144
prime-rl42.8no readingno reading2,224
ROLL0.714no reading3,448
Search-R10246no reading5,568
Trinity-RFT0.51no reading712
TRL57.8no reading15819,683
Unsloth796.2no reading2679,133
verl25.2no readingno reading24,157

Evidaxis snapshot · period 2026-11-07 · n = 16 systems · method medians of measured values · updated · version 868e3c39d9c7-b1 · CSV for this niche

What changed since the last snapshot?

3 published readings changed between snapshots d1eb0cec0f6a and 868e3c39d9c7.

slope chart: Trinity-RFT, changed published values, 2, 2026-10-31 to 2026-11-07
2 published values changed for Trinity-RFT since 2026-10-31. Evidaxis snapshots · .
Snapshot values
MetricPreviousCurrentDelta
Commits per week0.60.5-0.1
Citing worksno reading1no reading
Weekly direct dependentsno readingno readingno reading
Stars7097123

Evidaxis snapshots · period 2026-10-31 to 2026-11-07 · n = 21 partitions · method paired snapshot values · updated · version 868e3c39d9c7-b1 · History CSV

Take the data or follow changes

Values and captured history are available under CC0-1.0. The system feed records changes to published numbers.

Values CSVHistory CSVJSON recordAtom feedJSON FeedAdd Evidaxis as a preferred source on GoogleReport an error

Evidaxis · period 2026-11-07 · n = 1 system · method CC0-1.0 open export · updated · version 868e3c39d9c7-b1

For teams tracking many systems

Use the cohort CSV, snapshot archive and membership manifest to reproduce observations across systems.

Cohort CSV · Snapshot archive · Membership manifest

Paper and links

Source records identify the system, repository and linked works; no paper authors are published.

What is not measured for Trinity-RFT, and why

Missing readings describe instrument coverage. A missing value is not a measured zero.

Unmeasured slots
ReadingReason
Weekly direct dependentsno partition

Direct dependents: no reading captured in partition not captured; the value is missing, not zero.

Gate ETA: never published; Evidaxis makes no public actionable prediction.

Dependents ecosystem composition: no reading; this field is not present in the snapshot.

Evidaxis · period 2026-11-07 · n = 1 system · method measurement coverage · updated · version 868e3c39d9c7-b1

Measurement detail

Gate, gauges, percentile, derived signals and snapshot receipt

History: 19 weekly observations; minimum 4. Axis coverage: 1 of 3 axes measurable. Gate: not computable - 1 of 3 axes measurable

Gate: not computable - 1 of 3 axes measurable

Trinity-RFT

repo
Post-training & Alignment · n = 16 · agentscope-ai/Trinity-RFT
1 of 3 axes measurable · conf low
convergence gate
Convergence gate: Gate: not computable - 1 of 3 axes measurable Development velocity: measured × Citation momentum: not measurable ∅ Direct-dependents momentum (deps.dev): not measurable ∅
gate
0 of 3 axes

Gate: not computable - 1 of 3 axes measurable

· Evidaxis snapshot · 2026-11-07
Gate: not computable - 1 of 3 axes measurable
position in cohort distribution one of 16 in cohort
percentile 7 within cohort p25 p50 p75 p0 p100
percentile
p7
· Evidaxis snapshot · 2026-11-07
momentum averages present-axis z-scores; two- and three-axis scores are not strictly comparable
velocity z
velocity z gaugevelocity z-2.19 within cohort, log-slope -0.088/wk.0-2.19
velocity z
-2.19 σ
slope
-0.088/wk
· Evidaxis snapshot · 2026-11-07
log-slope -0.088/wk
citation z Citation axis: insufficient source coverage (source returned too few observations).
direct-dependents z Direct-dependents axis: minimum of 14 usable points of the new quantity and 5 latest dependents not met.
cohort
16
systems measured
weekly commits
0.5
Development velocity axis: measured.
total citations
1
Citation axis: insufficient source coverage (source returned too few observations).
momentum
22.7
percentile p7
commit heatmap, last 26 weeks0.5/wk
weekly commits, 26 weeks26 weeks of commits, peak 11 in a week, 16 active weeks, 64 total.week 1: 4 commitsweek 2: 7 commitsweek 3: 11 commitsweek 4: 10 commitsweek 5: 3 commitsweek 6: 7 commitsweek 7: 7 commitsweek 8: 2 commitsweek 9: 1 commitsweek 10: 1 commitsweek 11: 3 commitsweek 12: 1 commitsweek 13: 0 commitsweek 14: 1 commitsweek 15: 0 commitsweek 16: 0 commitsweek 17: 0 commitsweek 18: 1 commitsweek 19: 0 commitsweek 20: 0 commitsweek 21: 4 commitsweek 22: 1 commitsweek 23: 0 commitsweek 24: 0 commitsweek 25: 0 commitsweek 26: 0 commits
weeks
26
peak/wk
11
active
16
total
64
· Evidaxis snapshot · 2026-11-07
citations by yearCitation axis: insufficient source coverage (source returned too few observations).
Daily dependents count (unscored) 0 dependents 0 direct · pypi/trinity-rft · deps.dev · the scored m5 axis uses weekly partitions
recency-weighted swing trailing 13-week vs prior 39-week log-slope ratio +0.43, regime recent slope below prior window. prior slope -0.032, recent slope -0.014 per week. -0.032prior 39w-0.014recent 13w+0.43×
swing
+0.43×
regime
recent slope below prior window
windows
13w vs 39w
snapshot 1 of 1 · · m5
commit-axis Sharpe log-linear slope over residual std, value -0.062 with 95% CI -0.090 to -0.035, 50 degrees of freedom. CI spans zero.0-0.06295% CI -0.090, -0.035-0.090-0.035
sharpe
-0.062
95% ci
-0.090, -0.035
dof
50
snapshot 1 of 1 · · m5
commit changepoint52 weekly commit points. No significant changepoint detected (|delta slope| within the noise floor).no significant changepoint
changepoint
none detected
snapshot 1 of 1 · · m5
code-leads-citation lag reserved

reserved: citation series too short / axis absent for cross-correlation

status
reserved
value
insufficient data
snapshot 1 of 1 · · m5
axis alpha vs cohort reserved

reserved: cohort N too small for regression (power ~0)

status
reserved
value
insufficient data
snapshot 1 of 1 · · m5
methodology gate snapshot 1 of 1 · · m5
commit weeks
52
citation years
1
shipped
3
reserved
3

shipped

  • recency-swing
  • commit-sharpe
  • commit-changepoint

reserved

  • code-leads-citation-lag
  • axis-alpha
  • gate-eta
snapshot 21 of 21 · · m5

Direct-dependents momentum (deps.dev)

Direct-dependents axis: minimum of 14 usable points of the new quantity and 5 latest dependents not met.

Coverage state
below_floor
Cohort z
no axis
OLS slope
no reading
Theil-Sen slope
no reading
Latest direct dependents
no reading
Confirmed-clean points
0
Reconstructed history points
no reading
As-of partition
no reading
Fragility veto
no reading
Rising vote
no

Source: deps.dev (Google), CC-BY 4.0. Modified: aggregated to system-level counts and momentum scores.

Axis definition · Snapshot axis correlations

Peer signals
SystemMomentumPercentile
AReaL58.573
ART54.660
Axolotl56.667
LLaMA-Factory15.60
ms-swift40.327
NeMo-RL60.580
Open R141.140
Open-Instruct69.593
OpenRLHF40.433
prime-rl61.387
ROLL25.213
Search-R143.947
Trinity-RFT22.77
TRL51.953
Unsloth81.8100
verl31.420

Evidaxis · period 2026-11-07 · n = 1 system · method measurement detail · updated · version 868e3c39d9c7-b1 · record JSON · snapshot receipt