Symmetry under affine reparameterizations of hidden coordinates selects a unique hierarchy of shallow coordinate-stable probes and a probe-visible quotient for cross-model transfer.
hub
and Gardner, Matt and Belinkov, Yonatan and Peters, Matthew E
11 Pith papers cite this work, alongside 134 external citations. Polarity classification is still indexing.
11
Pith papers citing it
134
external citations · Crossref
hub tools
citation-role summary
background 1
citation-polarity summary
verdicts
UNVERDICTED 11roles
background 1polarities
background 1representative citing papers
Neurons exhibit concept-conditioned activation ranges forming Gaussian-like distributions with minimal overlap, and range-based interventions via NeuronLens outperform neuron-level masking in targeted manipulation with reduced collateral effects.
LLMs recover dominant binomial orders from corpora but align less closely with exact preference distributions, with preference strength partially encoded in middle-to-late layers and manipulable via steering.
Introduces LOES, a constructive spectral method to select task-discriminative subspaces from intermediate layer embeddings, and GeoReg for enforcing simplicial class geometry during fine-tuning, with reported gains increasing with model depth across modalities.
LLMs represent semantic relations geometrically via embedding distance and direction; a linear Polar Probe decodes these structures from middle-layer activations and generalizes to new entities.
Larger LLMs reproduce constructional productivity via entrenchment in coercion cases with nonce words but fail to use statistical preemption to avoid overgeneralizing semantically plausible but unobserved patterns.
Trait-space drift monitoring detects emergent misalignment checkpoints in 7-9B LLMs with 2.2% FNR, 2.9% FPR and 0.99 AUROC, outperforming PCA and SAE baselines.
MIPIC trains Matryoshka representations using self-distilled intra-relational alignment and progressive information chaining, yielding competitive results on STS, NLI, and classification tasks especially at low dimensions.
Inflectional features stay linearly decodable across all layers while lexical identity weakens with depth in modern transformers.
LRP-based attention head selection and distributed application improve the efficiency and accuracy of function vectors for steering LLMs compared to prior choices.
Probing classifiers are a common but limited method for analyzing linguistic knowledge in neural NLP models, and this review outlines their promises, methodological shortcomings, and recent advances.
citing papers explorer
-
Deep Minds and Shallow Probes
Symmetry under affine reparameterizations of hidden coordinates selects a unique hierarchy of shallow coordinate-stable probes and a probe-visible quotient for cross-model transfer.
-
Neurons Speak in Ranges: Breaking Free from Discrete Neuronal Attribution
Neurons exhibit concept-conditioned activation ranges forming Gaussian-like distributions with minimal overlap, and range-based interventions via NeuronLens outperform neuron-level masking in targeted manipulation with reduced collateral effects.
-
Behavioral and Representational Evidence of Binomial Ordering Preferences in Large Language Models
LLMs recover dominant binomial orders from corpora but align less closely with exact preference distributions, with preference strength partially encoded in middle-to-late layers and manipulable via steering.
-
Uncovering the Latent Potential of Deep Intermediate Representations
Introduces LOES, a constructive spectral method to select task-discriminative subspaces from intermediate layer embeddings, and GeoReg for enforcing simplicial class geometry during fine-tuning, with reported gains increasing with model depth across modalities.
-
Polar probe linearly decodes semantic structures from LLMs
LLMs represent semantic relations geometrically via embedding distance and direction; a linear Polar Probe decodes these structures from middle-layer activations and generalizes to new entities.
-
Linguistic Productivity in Large Language Models: Models Coerce, but do not Preempt
Larger LLMs reproduce constructional productivity via entrenchment in coercion cases with nonce words but fail to use statistical preemption to avoid overgeneralizing semantically plausible but unobserved patterns.
-
Trait-space Monitoring for Emergent Misalignment During Supervised Finetuning
Trait-space drift monitoring detects emergent misalignment checkpoints in 7-9B LLMs with 2.2% FNR, 2.9% FPR and 0.99 AUROC, outperforming PCA and SAE baselines.
-
MIPIC: Matryoshka Representation Learning via Self-Distilled Intra-Relational and Progressive Information Chaining
MIPIC trains Matryoshka representations using self-distilled intra-relational alignment and progressive information chaining, yielding competitive results on STS, NLI, and classification tasks especially at low dimensions.
-
Model Internal Sleuthing: Finding Lexical Identity and Inflectional Features in Modern Language Models
Inflectional features stay linearly decodable across all layers while lexical identity weakens with depth in modern transformers.
-
Fast & Faithful Function Vectors
LRP-based attention head selection and distributed application improve the efficiency and accuracy of function vectors for steering LLMs compared to prior choices.
-
Probing Classifiers: Promises, Shortcomings, and Advances
Probing classifiers are a common but limited method for analyzing linguistic knowledge in neural NLP models, and this review outlines their promises, methodological shortcomings, and recent advances.