The TwitterAAE corpus of [Blodgett et al. (2016)](https://aclanthology.org/D16-1120/) for measuring language model performance on African-American-aligned Tweets.

twitter_aae_aa

The TwitterAAE corpus of [Blodgett et al. (2016)](https://aclanthology.org/D16-1120/) for measuring language model performance on African-American-aligned Tweets.

Bits/byte: Average number of bits per byte according to model probabilities.

The TwitterAAE corpus of [Blodgett et al. (2016)](https://aclanthology.org/D16-1120/) for measuring language model performance on African-American-aligned Tweets.

Denoised inference runtime (s): Average time to process a request to the model minus performance contention by using profiled runtimes from multiple trials of SyntheticEfficiencyScenario.

The TwitterAAE corpus of [Blodgett et al. (2016)](https://aclanthology.org/D16-1120/) for measuring language model performance on African-American-aligned Tweets.

# eval: Number of evaluation instances.

The TwitterAAE corpus of [Blodgett et al. (2016)](https://aclanthology.org/D16-1120/) for measuring language model performance on African-American-aligned Tweets.

# train: Number of training instances (e.g., in-context examples).

The TwitterAAE corpus of [Blodgett et al. (2016)](https://aclanthology.org/D16-1120/) for measuring language model performance on African-American-aligned Tweets.

truncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).

The TwitterAAE corpus of [Blodgett et al. (2016)](https://aclanthology.org/D16-1120/) for measuring language model performance on African-American-aligned Tweets.

# prompt tokens: Number of tokens in the prompt.

The TwitterAAE corpus of [Blodgett et al. (2016)](https://aclanthology.org/D16-1120/) for measuring language model performance on African-American-aligned Tweets.

# output tokens: Actual number of output tokens.

The TwitterAAE corpus of [Blodgett et al. (2016)](https://aclanthology.org/D16-1120/) for measuring language model performance on African-American-aligned Tweets.

# trials: Number of trials, where in each trial we choose an independent, random set of training instances.

twitter_aae_aa_demographic:aa

min=2.53, mean=2.53, max=2.53, sum=2.53 (1)

min=0.263, mean=0.263, max=0.263, sum=0.263 (1)

min=1000, mean=1000, max=1000, sum=1000 (1)

min=11.764, mean=11.764, max=11.764, sum=11.764 (1)

min=2.603, mean=2.603, max=2.603, sum=2.603 (1)

min=0.24, mean=0.24, max=0.24, sum=0.24 (1)

min=2.559, mean=2.559, max=2.559, sum=2.559 (1)

min=0.287, mean=0.287, max=0.287, sum=0.287 (1)

min=2.489, mean=2.489, max=2.489, sum=2.489 (1)

min=2.443, mean=2.443, max=2.443, sum=2.443 (1)

min=2.462, mean=2.462, max=2.462, sum=2.462 (1)

min=2.518, mean=2.518, max=2.518, sum=2.518 (1)

min=2.359, mean=2.359, max=2.359, sum=2.359 (1)

min=0.731, mean=0.731, max=0.731, sum=0.731 (1)

min=14.861, mean=14.861, max=14.861, sum=14.861 (1)

min=5.874, mean=5.874, max=5.874, sum=5.874 (1)

min=2.328, mean=2.328, max=2.328, sum=2.328 (1)

min=0.035, mean=0.035, max=0.035, sum=0.035 (1)

min=13.993, mean=13.993, max=13.993, sum=13.993 (1)

min=2.669, mean=2.669, max=2.669, sum=2.669 (1)

min=0.319, mean=0.319, max=0.319, sum=0.319 (1)

min=14.845, mean=14.845, max=14.845, sum=14.845 (1)

min=2.693, mean=2.693, max=2.693, sum=2.693 (1)

min=0.204, mean=0.204, max=0.204, sum=0.204 (1)

min=2.715, mean=2.715, max=2.715, sum=2.715 (1)

min=0.189, mean=0.189, max=0.189, sum=0.189 (1)

min=2.913, mean=2.913, max=2.913, sum=2.913 (1)

min=0.215, mean=0.215, max=0.215, sum=0.215 (1)

min=2.694, mean=2.694, max=2.694, sum=2.694 (1)

min=2.817, mean=2.817, max=2.817, sum=2.817 (1)

min=3.019, mean=3.019, max=3.019, sum=3.019 (1)

min=1.003, mean=1.003, max=1.003, sum=1.003 (1)

min=3.151, mean=3.151, max=3.151, sum=3.151 (1)

min=2.302, mean=2.302, max=2.302, sum=2.302 (1)

min=0.438, mean=0.438, max=0.438, sum=0.438 (1)

min=2.254, mean=2.254, max=2.254, sum=2.254 (1)

min=0.824, mean=0.824, max=0.824, sum=0.824 (1)

min=14.966, mean=14.966, max=14.966, sum=14.966 (1)

min=2.114, mean=2.114, max=2.114, sum=2.114 (1)

min=0.134, mean=0.134, max=0.134, sum=0.134 (1)

min=2.13, mean=2.13, max=2.13, sum=2.13 (1)

min=1.285, mean=1.285, max=1.285, sum=1.285 (1)

min=2.587, mean=2.587, max=2.587, sum=2.587 (1)

min=2.673, mean=2.673, max=2.673, sum=2.673 (1)

min=2.476, mean=2.476, max=2.476, sum=2.476 (1)

min=0.185, mean=0.185, max=0.185, sum=0.185 (1)

min=2.556, mean=2.556, max=2.556, sum=2.556 (1)

min=0.08, mean=0.08, max=0.08, sum=0.08 (1)

min=2.646, mean=2.646, max=2.646, sum=2.646 (1)

min=0.111, mean=0.111, max=0.111, sum=0.111 (1)

min=2.764, mean=2.764, max=2.764, sum=2.764 (1)

min=0.136, mean=0.136, max=0.136, sum=0.136 (1)

min=2.522, mean=2.522, max=2.522, sum=2.522 (1)

min=2.478, mean=2.478, max=2.478, sum=2.478 (1)

min=0.159, mean=0.159, max=0.159, sum=0.159 (1)

min=3.034, mean=3.034, max=3.034, sum=3.034 (1)

min=0.12, mean=0.12, max=0.12, sum=0.12 (1)

min=3.385, mean=3.385, max=3.385, sum=3.385 (1)

min=0.122, mean=0.122, max=0.122, sum=0.122 (1)

min=3.963, mean=3.963, max=3.963, sum=3.963 (1)

min=0.077, mean=0.077, max=0.077, sum=0.077 (1)