That depends on what kind of bridging you want to do. Most token ring applications use source route bridging. To convert from that to ethernet format, you require the source bridge translational function which requires the IBM feature set.
The following URL describes how to configure this feature
http://www.cisco.com/warp/customer/100/48.shtml
You can configure transparant bridging between ethernet and token ring with the IP only feature set, but it does have some limitations due the the way that the addresses are put onto the lans in reversed format when going from token ring to ethernet and vice versa.
This means that any frames that contain mac addresses in the data portion, such as ARP frames, will receive them in the wrong format.
The bitswap-layer3-address parameter on the brigde-group command was introduced specifically to bitswap this address but for ARP frames only. Other protocols, such as DHCP, will not work.