IC-1456Pre-trained language models fail to predict both interpretations of ambiguous inputs in zero-shot semantic parsing

Elias Stengel-Eskin, Kyle Rawlins, Benjamin Van Durme

SourceZero and Few-shot Semantic Parsing with Ambiguous Inputs

The paper tests Codegen (350M, 2B, 6B, 16B), Llama-13B, Vicuna-13B, and GPT-3.5-turbo on 200 examples per ambiguity type (PP attachment, scope, inverse scope, bound pronoun, conjunction) in a zero-shot setting where the prompt provides the compositional ingredients for both derivations but no examples of the ambiguity type. Most models score 0.00 on ZM5 (both LFs present in top-5 beam search outputs) across all five types. GPT-3.5-turbo achieves high single-LF accuracy on PP and scope but still scores 0.00 on ZM5. Llama-13B scores 0.00 on every type, while its instruction-tuned variant Vicuna-13B reaches 4.00 on bound and 9.50 on conjunction. Token-level confidence for Codegen-16B partially reflects the divergence point for scope and inverse scope but not for PP attachment or bound pronouns.

Evidence
correlational
Key metric
ZM5 (pp/scope/revscope/bound/conj): cg-350m 0.00/0.00/0.00/0.00/1.00, cg-2b 0.00/0.00/0.00/0.50/0.00, cg-6b 0.00/0.00/0.00/0.00/3.50, cg-16b 0.00/0.00/0.00/3.50/15.00, ll-13b 0.00/0.00/0.00/0.00/0.00, vi-13b 0.00/0.00/0.00/4.00/9.50, gpt-3.5 0.00/0.00/0.00/0.00/0.00
Caveat
The paper notes that decoding strategies emphasizing diversity (temperature sampling, sequential decoding) might improve ZM scores, and that the task is a compositional generalization challenge requiring the model to combine structures into novel derivations.
Model
CodeGen, LLaMA, Vicuna, GPT-3.5 / ChatGPT-3.5 GPT-3.5-turbo
Concepts
Failure mode
Methods
Constrained Decoding [primary], BenchCLAMP [supporting]
Related findings
IC-1457
Extraction
automatic-extraction