Pith. sign in

REVIEW 9 cited by

Designing and Interpreting Probes with Control Tasks

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 1909.03368 v1 pith:G7P6DUPK submitted 2019-09-08 cs.CL

classification cs.CL
keywords linguisticprobetasksaccuracycontrolprobestaskelmo
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

Probes, supervised models trained to predict properties (like parts-of-speech) from representations (like ELMo), have achieved high accuracy on a range of linguistic tasks. But does this mean that the representations encode linguistic structure or just that the probe has learned the linguistic task? In this paper, we propose control tasks, which associate word types with random outputs, to complement linguistic tasks. By construction, these tasks can only be learned by the probe itself. So a good probe, (one that reflects the representation), should be selective, achieving high linguistic task accuracy and low control task accuracy. The selectivity of a probe puts linguistic task accuracy in context with the probe's capacity to memorize from word types. We construct control tasks for English part-of-speech tagging and dependency edge prediction, and show that popular probes on ELMo representations are not selective. We also find that dropout, commonly used to control probe complexity, is ineffective for improving selectivity of MLPs, but that other forms of regularization are effective. Finally, we find that while probes on the first layer of ELMo yield slightly better part-of-speech tagging accuracy than the second, probes on the second layer are substantially more selective, which raises the question of which layer better represents parts-of-speech.

Discussion (0). Sign in to comment.

Forward citations

Cited by 9 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. OpenAlex reports about 12 citations worldwide. Full citation record

  1. Robust Control under Stationary Ambiguity

    cs.LG 2026-08 conditional novelty 7.0 of 10

    Policies trained under stationary latent ambiguity, implemented by refreshing the latent parameter, preserve robustness to regime shifts better than policies trained under a fixed latent draw.

  2. Verbalizable Representations Form a Global Workspace in Language Models

    cs.CL 2026-07 conditional novelty 7.0 of 10

    Language models represent their current reasoning in a small, readable set of verbalizable vectors (the J-space) that functions like a global workspace.

  3. When Does Reward Teach State? A Hidden-Automaton Instrument and the Group-Language Boundary

    cs.LG 2026-07 conditional novelty 7.0 of 10

    High reward in sparse RL does not imply latent-state recovery; a hidden-DFA instrument separates perception from planning gaps and flags group-language structure as a pre-training warning.

  4. Two Stages of Folding: Convergent Mechanisms in AI Protein Folding Trunks

    cs.LG 2026-02 conditional novelty 7.0 of 10

    Causal interventions show ESMFold's folding trunk first transfers sequence chemistry into its pairwise representation (blocks 0–7), then builds pairwise spatial features that control output geometry (blocks 25+).

  5. Mechanistic Interpretability of Cognitive Complexity in LLMs via Linear Probing using Bloom's Taxonomy

    cs.AI 2026-02 conditional novelty 6.0 of 10

    Linear probes on LLM residual streams classify Bloom's Taxonomy levels with high accuracy, but the result may reflect prompt lexico-semantic cues rather than a general cognitive-complexity representation.

  6. Quantitative Analysis of Proxy Tasks for Anomalous Sound Detection

    eess.AS 2026-01 conditional novelty 6.0 of 10

    Better proxy-task performance does not generally improve anomalous sound detection; only source separation showed a strong, consistent positive correlation.

  7. Tracking World States with Language Models: State-Based Evaluation Using Chess

    cs.AI 2025-08 conditional novelty 6.0 of 10

    A model-agnostic chess evaluation metric measures state-tracking fidelity by comparing legal-move sets of predicted and true positions, showing GPT-4o's reconstruction quality degrades over longer games.

  8. What Does it Mean for a Neural Network to Learn a "World Model"?

    cs.AI 2025-07 conditional novelty 6.0 of 10

    Defines a world model as a simple commutative-diagram factorization through an intermediate representation, with conditions that the model be learned and emergent rather than inherited from input or output.

  9. Towards Transparent AI: A Survey on Explainable Large Language Models

    cs.CL 2025-06 conditional novelty 3.0 of 10

    A review that groups LLM explainability methods by transformer architecture and discusses their evaluation and applications.

Pith tools