History Forcing: Distilling History-Dependent Continuations for Long-Horizon Video World Models
History Forcing is a distillation framework for few-step autoregressive video diffusion models that strengthens historical supervision.
Source-based brief · Editor review not recorded · See checks and limits ↓
One primary source; other links are context, not independent corroboration.
- Primary sourceHugging Face
- Discovery linkHugging Face
The paper page is linked. Peer review and findings have not been independently checked by AINewsIQ.
Open full article view →- Original paper+45
- Fetched source page+30
- No editor link review+0
- Original non-discussion link+5
This measures the stored source trail and editorial steps. It does not certify accuracy, peer review, or independent corroboration. Popularity adds no points.
View full scoring →Read the scoring rubric →What the source says
History Forcing is a distillation framework for few-step autoregressive video diffusion models that strengthens historical supervision. It computes the residual between teacher predictions with and without historical context, weights it, and adds it to the distillation target to enforce long-range consistency. The authors report a VBench quality score of 85.07 for 15s video and a VBench-Long score of 79.82 for 60s generation for 1.3B models. For out-of-sight world modeling, they report Evolution Quality rising from 2.35 to 2.70 on PlayWorld and Object Consistency from 58.43 to 62.18 on MemoBench.
Claim explorer
Attribution, source and conditionsSource-stated points from the stored brief. Claim-level assessments have not been recorded.
Source claims not independently checked; benchmark results are self-reported by the authors.
This connects a source, an attributed statement and its qualifications. It does not establish that the claim is correct.
Questions worth asking
Open questions for further reportingWhat remains unconfirmed?
Source claims not independently checked; benchmark results are self-reported by the authors.
What material was available for this brief?
The recorded brief basis is fetched source page. These points are AINewsIQ paraphrases, not quotations or independent confirmations.
The method uses the residual between teacher predictions with and without historical context to decouple supervision.
Source: Hugging Face ↗Stored source-based brief · key point 1 · AINewsIQ paraphraseReported VBench quality is 85.07 for 15s video generation and VBench-Long is 79.82 for 60s generation.
Source: Hugging Face ↗Stored source-based brief · key point 2 · AINewsIQ paraphraseOn PlayWorld, Evolution Quality improves from 2.35 to 2.70; on MemoBench, Object Consistency improves from 58.43 to 62.18.
Source: Hugging Face ↗Stored source-based brief · key point 3 · AINewsIQ paraphraseLimits from the stored brief
Source claims not independently checked; benchmark results are self-reported by the authors.
This is a source-based abstract, not independent validation of the source's claims.
Evidence & sources
These links show where the signal came from and what the source page linked to. A linked page is context, not independent corroboration.
The paper page is linked. Peer review and findings have not been independently checked by AINewsIQ.
Timeline & updates
- Signal dated by source or discovery feed
This timestamp places the signal in the radar. It may be the discovery post date rather than the primary page's first publication date.
- Shared brief stored
AINewsIQ prepared one abstract from fetched source page.
Brief history · 1 retained version
These are changes to AINewsIQ’s own publication. Earlier history is unavailable unless recorded here. Complete retained versions are available to the editor.
- Version 1 · published
Source-based brief stored or updated. This does not establish that the publisher changed its page.
Automated record
Source observations · 1 recent captures
A changed capture means the extracted text differed. It can reflect page furniture or extraction changes; it does not establish a substantive correction. Source captures remain private.
Retained capture; no earlier comparison available.
New engagement snapshots do not mean the source itself changed. Source observations and AINewsIQ brief revisions are separate records.
Why this may matter
Long video generation often loses consistency when subjects leave and reappear in view. A distillation method that injects historical constraints could make few-step autoregressive video world models more usable for longer rollouts.
Limits and open questions
Source claims not independently checked; benchmark results are self-reported by the authors.
Organization-specific financial, security, and operational effects require the organization's own context and review.
Join the discussion
Have a different perspective or a question about the evidence? Discussion is free with a verified email account. Every comment is reviewed before publication.
No published comments yet.
New commenters: 2 submissions per 24 hours, at least 5 minutes apart, until 3 are approved. Discussion guidelines and moderation →