Zenaique

Predict which response wins after length normalization is applied

Predict output·Medium·4.0 · 0·~2 min·Asked atBcgNiki AiNotion
Attempt it
Two candidates answer the same prompt. Response A: raw reward 6.0, 300 tokens. Response B: raw reward 4.5, 90 tokens. Under SimPO style length normalization (reward divided by token count), which response is preferred, and what are the two normalized scores?
TL;DR

B wins because reward density (4.5 / 90 = 0.05) beats A's diluted density (6.0 / 300 = 0.02), even though A's raw score is higher.

Memory aid
Sign in to see the mnemonic that makes this stick.
Easy to grasp

Imagine grading two essays. The first got 60 points but is 300 sentences long. The second got 45 points in only 90 sentences. If you ask which is better overall, you might say the first. But if you ask which earned more points per sentence, the second is way ahead, 0.5 versus 0.2. Length normalization is just that per-sentence view applied to AI responses. It stops the model from winning by piling on words. The math here is the same idea: divide each score by length and the shorter, denser answer wins.

Key concepts

Concept explanation~2 min read

Everything you need to truly understand this topic: intuition, mechanics, step by step explanation, code, formulas, and worked example. Click to expand.

This is a short arithmetic question wrapped around a long structural point. The numbers are easy: 0.02 versus 0.05, B wins. The interesting work is understanding why a single division operation is enough to dissolve one of the largest reward-hacking patterns in RLHF, and what new failure modes it introduces.

Length bias in RLHF arises because human annotators tend to prefer longer, more detailed-looking answers. The reward model absorbs that correlation. During PPO or DPO, the policy exploits it by padding. The remedy that emerged in 2024 and consolidated in 2026 practice is to score per token rather than per sequence, formalized in SimPO.

This deep dive walks through the arithmetic, the gradient geometry change that makes it work, the practical evaluation lever (length-controlled win rate), and the new failure modes that a length-normalized objective opens up. The goal is to make density-based scoring a tool you can defend and tune, not a magic word.

Doing the arithmetic and what it means

Reward density is just reward divided by length. For Response A, that is 6.0 / 300 = 0.02 per token. For Response B, 4.5 / 90 = 0.05 per token. B is 2.5x denser than A on this metric.

The raw versus density flip is the whole pedagogical point. A's raw advantage of 1.5 points looked decisive in absolute terms. Once you charge per token, A's extra 210 tokens become a cost rather than a free bonus, and the comparison inverts. The threshold for the flip is set by the length ratio: any time the longer response's reward to length ratio falls below the shorter response's, density wins.

A quick rule of thumb. If A is k times longer than B, A needs at least k times B's raw reward to beat B on density. Here A is 3.33x longer than B, so A would have needed at least 15.0 (3.33 * 4.5) to win. It only had 6.0. Density flips the call cleanly.

Do the arithmetic. Response A has total reward rA=2.4r_A = 2.4 over LA=60L_A = 60 tokens. Response B has rB=3.6r_B = 3.6 over LB=120L_B = 120 tokens. Raw reward picks B (3.6 > 2.4). Length-normalized reward picks A: rA/LA=0.040r_A / L_A = 0.040 vs rB/LB=0.030r_B / L_B = 0.030. A is denser. The normalization changes the answer because it asks not which response scored higher overall, but which had more value per unit length. That is the operational definition of conciseness as a first-class objective rather than a side constraint.

How SimPO uses this in the training objective
Length-controlled evaluation and why it matters
New failure modes density opens up
Sign in to unlock the full deep dive.

Situations where this technique stops working.

Sign in to see when this approach fails.

2–4 min · Everything important, quickly.

Sign in to see the quick scan of the deep dive.

Real products, models, and research that use this idea.

  • SimPO uses average per-token log-likelihood as the implicit reward, eliminating the length sensitivity DPO inherits from sequence-level scoring
  • AlpacaEval's length-controlled (LC) win rate normalizes for response length before declaring a winner, and several Llama 3 variants jumped notably in LC ranking versus raw
Sign in to see more production examples.

What an interviewer would ask next. Try answering before peeking at the approach.

QWhy does DPO show length bias while SimPO does not, even with the same preference data?
A

Compare the two loss formulas at the token level. DPO compares sequence log-likelihoods so longer chosen responses move the loss more; SimPO averages per token so length cancels in the gradient.

1 more follow-up an interviewer would ask next. Sign in to reveal them.

Red flags & common mistakes

The phrases that signal junior thinking. Click to expand.

Most common mistake

Comparing raw rewards directly and concluding A wins. Length normalization specifically inverts that comparison when length differences are large.

Sign in to see all red flags and common mistakes.

60 second bullets to scan on the way to the call.

  • How to compute reward density given raw reward and token count

  • Why SimPO uses per-token average instead of sequence log-likelihood

Sign in to unlock the revision sheet.

Primary sources. Browse if you want the original framing.

Similar questions

Same topic, related formats. Practice these next.

4 curated
Next question
What is RLHF, and why is it used after pretraining?
MCQ·Easy