This 2018 BBC video says about itself:
Steve Backshall looks at the Majungasaurus, the only know dinosaur to eat others from its own species.
From PLOS ONE, 27 May 2020:
High frequencies of theropod bite marks provide evidence for feeding, scavenging, and possible cannibalism in a stressed Late Jurassic ecosystem
Bite marks provide direct evidence for trophic interactions and competition in the fossil record. However, variations in paleoecological dynamics, such as trophic relationships, feeding behavior, and food availability, govern the frequency of these traces.
Theropod bite marks are particularly rare, suggesting that members of this clade might not often focus on bone as a resource, instead preferentially targeting softer tissues.
Here, we present an unusually large sample of theropod bite marks from the Upper Jurassic Mygatt-Moore Quarry (MMQ). We surveyed 2,368 vertebrate fossils from MMQ in this analysis, with 684 specimens (28.885% of the sample) preserving at least one theropod bite mark.
This is substantially higher than in other dinosaur-dominated assemblages, including contemporaneous localities from the Morrison Formation. Observed bite marks include punctures, scores, furrows, pits, and striations. Striated marks are particularly useful, diagnostic traces generated by the denticles of ziphodont teeth, because the spacing of these features can be used to provide minimum estimates of trace maker size.
In the MMQ assemblage, most of the striations are consistent with denticles of the two largest predators known from the site: Allosaurus and Ceratosaurus. One of the bite marks suggests that a substantially larger theropod was possibly present at the site and are consistent with large theropods known from other Morrison Formation assemblages (either an unusually large Allosaurus or a separate, large-bodied taxon such as Saurophaganax or Torvosaurus).
The distribution of the bite marks on skeletal elements, particularly those found on other theropods, suggest that they potentially preserve evidence of scavenging, rather than active predation. Given the relative abundances of the MMQ carnivores, partnered with the size-estimates based on the striated bite marks, the feeding trace assemblage likely preserves the first evidence of cannibalism in Allosaurus.
See also here.