Paper Citation Record · LEDGER

The WER Trap: Shattering the Illusion of Unified Tokens in Speech Language Models

As of 10 August 2026, this Paper Citation Record lists 3 of 3 outbound references and 1 inbound Pith citation observation for arXiv:2605.29209.

A citation records a reference. It does not transfer a finding from one paper to another.

pith.paper-citation-record.v1
2605.29209 v1

Coverage vector

measured 3 of 3 reference resolution

Typed states for the displayed outbound observations.

Source: paper_references, paper_reference_links, observed 2026-06-29T06:02:07.932691Z

measured 4 of 4 standing notices

One-hop event checks from named stored sources.

Source: scholarly_work_events, retraction_status_cache, observed 2026-08-10T06:31:04.303077+00:00

measured 1 of 1 inbound itemization

Pith citing papers itemized under the disclosed page cap.

Source: paper_references, paper_reference_links, observed 2026-08-03T10:13:57.787394Z

measured 0 of 1 external citation measurements

A source-named dated measurement, never combined with another source.

Source: cited_works

Reference resolution

3 of 3 outbound references displayed

verified exact1
verified fuzzy0
unresolved1
parse uncertain0
malformed identifier0
metadata mismatch1

External citation measurements

No source-named external measurement is stored.

Outbound references

Observation 5f8f2a8e-d722-4b33-879c-17de4a9412ca · outbound

This paper cites MMAU: A Massive Multi-Task Audio Understanding and Reasoning Benchmark.

The WER Trap: Shattering the Illusion of Unified Tokens in Speech Language Models MMAU: A Massive Multi-Task Audio Understanding and Reasoning Benchmark

Reference 1

Resolution: verified exact; local_arxiv, observed 2026-06-29T06:03:08.411387Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-10T06:31:04.303077+00:00.

Observation 540b4a8b-dd47-466f-b4b3-ec4f989f556b · outbound

This paper cites Step-audio-r1 technical report.

The WER Trap: Shattering the Illusion of Unified Tokens in Speech Language Models Step-audio-r1 technical report

Reference 2

Resolution: metadata mismatch; arxiv_id, observed 2026-06-29T06:03:08.408192Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-10T06:31:04.303077+00:00.

Observation dc72c124-18db-419a-80a7-5dd153362632 · outbound

This paper cites Tokenizer Backbone.Both the encoder and decoder utilize a robust Transformer architecture: • Encoder:32 layers, 20 attention heads, 1280 hidden dimension, and 5120 linear units.

The WER Trap: Shattering the Illusion of Unified Tokens in Speech Language Models Tokenizer Backbone.Both the encoder and decoder utilize a robust Transformer architecture: • Encoder:32 layers, 20 attention heads, 1280 hidden dimension, and 5120 linear units

Reference 3

Resolution: unresolved; no resolver link, observed 2026-06-29T06:02:07.932691Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Pith citing papers

Observation 92418081-62ad-4080-9924-bd5911203b0b · inbound

ParaASR: Multi-Token Prediction for Fast and Long-Context LLM-Based Speech Recognition cites this paper.

ParaASR: Multi-Token Prediction for Fast and Long-Context LLM-Based Speech Recognition The WER Trap: Shattering the Illusion of Unified Tokens in Speech Language Models

Reference 39

Resolution: unresolved; no resolver link, observed 2026-08-03T10:13:57.787394Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.