Sure, so there's a few differences:

JUCE contains some "general" filter types

chowdsp_filters has these types of filters as well, but with a few differences:

FIR filter: our implementation uses SIMD (on Windows/Linux) and Accelerate (on Mac) to get better performance. To get a rough idea of the performance difference, you can take a look at the relevant benchmarks.
IIR Filter
- chowdsp::IIRFilter has a filter order that is fixed at compile-time
- Coefficient makers for basic filter types are accessible through sub-classes
- I believe our coefficient makers use some slightly different equations than the JUCE implementation
State Variable Filters: our implementation uses a slightly different derivation as well, and has a "multi-mode" mode.
Linkwitz-Riley Filter: our implementation covers more filter orders

This library also has some higher-order filter types:

chowdsp::NthOrderFilter: a higher-order filter based on State Variable Filters
chowdsp::ButterworthFilter: a traditional Nth-order Butterworth Filter
chowdsp::ChebyshevIIFilter: a traditional Nth-order type-II Chebyshev Filter
chowdsp::EllipticFilter: a traditional Nth-order Elliptic Filter

And then there's a few other odds and ends, which are probably not needed as often, but can be kind of fun:

chowdsp::ARPFilter: a linear emulation of the ARP 1047 multi-mode filter
chowdsp::WernerFilter: an implementation of the filter from this DAFx paper, but with a different derivation than the one in the paper.
chowdsp::FractionalOrderFilter: an approximate fraction-order filter based on this DAFx paper
chowdsp::ModFilterWrapper: wraps a chowdsp::IIRFilter as a state variable filter to allow for fast modulation

Filter differences between juce Butterworth filters #403

Description