Validator Weights

How validator weights express subnet miner evaluations before Yuma Consensus filters and aggregates them.

Validator weights are evaluation signals set by subnet validators. They express how validators assess miner work inside a subnet before Yuma Consensus filters and aggregates those signals into subnet outcomes (Yuma Consensus).

The term is subnet-specific. A validator weight only has meaning inside the task, scoring method, and incentive setting that produced it, so it is different from a general validator reputation score.

What Validator Weights Contain

A validator’s weights are carried as a weight vector: a list of miner targets paired with evaluation scores. The vector expresses relative standing among miners the validator evaluated, not an absolute performance grade detached from subnet context (Glossary: Weight Vector, Understanding Incentive Mechanisms).

Each entry names one miner the validator rated and the score assigned to that miner’s work. Validators can disagree on both which miners deserve attention and how high each score should be, so one validator’s vector is only one row of evidence before subnet-wide aggregation begins.

From One Vector to the Weight Matrix

Validators record weights through weight setting, the on-chain action that submits one validator’s vector for an epoch. Yuma Consensus processes the weight matrix, the subnet-level collection that stacks every participating validator’s row into structured consensus input (Glossary: Weight Matrix, Yuma Consensus).

The transaction-fee documentation lists validator weight submissions as fee-free for the base weight-and-length transaction fee. set_weights, commit_weights, and reveal_weights are named fee-free extrinsics, so their consensus role should not be read as carrying the same base fee path as ordinary fee-bearing state changes (Transaction Fees in Bittensor, Commit Reveal).

Reading down one miner column therefore gathers every validator’s score for that miner, while reading across one validator row recovers that validator’s full evaluation. Matrix vocabulary names the assembled input; validator-weights vocabulary names the per-validator scores inside each row (Glossary: Weight Matrix).

Stake Weight Scales Each Matrix Row

Validator weights and stake weight answer different questions on the same path. Validator weights are the per-miner scores a validator submits; stake weight is the stake-derived influence that determines how heavily that validator’s row counts when Yuma Consensus aggregates the matrix (Understanding Subnets: Validator stake weight, Yuma Consensus: Miner emissions).

Two validators can submit similar miner evaluations yet carry different influence when their stake weights differ. Yuma documentation describes miner-side aggregation as summing clipped validator weights scaled by validator stake, so matrix rows are stake-weighted inputs rather than one-vote-per-validator tallies (Yuma Consensus: Miner emissions).

That scaling sits between submission and outcomes. Validator weights name what a validator thought about miner work; stake weight names how much that opinion should move consensus relative to other validators on the selected subnet (Glossary: Stake Weight).

Distinction from Liquid Alpha

Validator weights are the evaluation signals validators submit for miners. Liquid alpha adjusts bond smoothing after those signals enter the consensus-based bond path (Glossary: Weight Vector, Yuma Consensus).

Weights vocabulary names the submitted scores; liquid alpha vocabulary names how bond history responds once consensus compares those scores on the validator-miner path.

Distinction from Weight Setting

Validator weights are the evaluation values inside a submission. Weight setting is the validator action that records those values on chain each epoch (Glossary: Weight Setting, Glossary: Weight Vector).

  • Validator weights — miner evaluation values in a submission.
  • Weight setting — validator action that records scores on chain.

Content vocabulary and submission-action vocabulary answer different validator stages (Yuma Consensus).

Distinction from Validator Trust

Validator weights are the evaluation signals a validator submits about miner work before clipping. Validator trust sums that validator’s clipped weight contributions across miners after the per-miner consensus benchmark filters outlier values (Glossary: Validator Trust, Glossary: Weight Vector, Yuma Consensus: Clipping).

  • Validator weights — evaluation signals submitted about miner work.
  • Validator trust — validator-side clipped influence summed across miners for one pass.

Submission vocabulary and post-clipping influence vocabulary answer different consensus layers (Yuma Consensus: Clipping).

Distinction from Consensus Score

Validator weights are the per-miner scores one validator records in a submission. Consensus score is the stake-weighted agreement benchmark Yuma compares those submissions against when clipping outlier values from the weight matrix (Glossary: Consensus Score, Yuma Consensus: Clipping).

  • Validator weights — one validator’s submitted miner scores before filtering.
  • Consensus score — per-miner benchmark that shapes which submitted values count.

Distinction from Trust

Validator weights are evaluation signals on the validator-input side of one submission. Trust is a miner-side ratio T_j = R_j / P_j read after the full matrix is clipped and aggregated, showing how much pre-rank support one miner retained (Glossary: Trust, Yuma Consensus: Miner emissions).

  • Validator weights — one validator’s submitted scores for miners.
  • Trust — miner-side clipping survival ratio after the matrix is processed.

Distinction from TAO Weight

TAO weight and validator weights answer different questions. TAO weight helps determine stake-derived influence when stake weight is formed, while validator weights express a validator’s miner evaluations inside Yuma Consensus (Glossary: TAO Weight, Glossary: Validator Weights, Yuma Consensus).

  • TAO weight — how TAO stake counts when forming stake weight.
  • Validator weights — the validator’s submitted miner evaluations.

Distinction from Yuma Consensus

Validator weights are the per-miner scores a validator records on chain through weight setting. Yuma Consensus settles those submissions into emission shares each tempo (Glossary: Weight Vector, Yuma Consensus).

  • Validator weights — submitted miner evaluation values before Yuma settlement.
  • Yuma Consensus — on-chain mechanism that uses validator weights to determine emission shares (Yuma Consensus: Validator emissions).

Distinction from Stake Weight

Validator weights are the per-miner scores a validator submits. Stake weight is the stake-derived influence applied when Yuma Consensus aggregates those scores in the matrix (Glossary: Weight Vector, Glossary: Stake Weight).

  • Validator weights — the per-miner scores a validator submits.
  • Stake weight — the stake-based influence applied to those scores in consensus.

Reader Boundary

Validator weights name subnet evaluation signals. They describe submitted miner-work assessments rather than miner leaderboards, reward guarantees, validator permits, or a complete explanation of Yuma Consensus (Yuma Consensus).

Specific examples belong to the subnet, mechanism, network, and source that produced the value at the time of observation.

Further Reading

Topics ConsensusValidation