Presents a new expert-curated dataset of multi-turn counterspeech dialogues in five languages targeting hate against seven groups, with span annotations linking to verified external knowledge for RAG applications.
hub
arXiv preprint arXiv:2304.14767 , year=
19 Pith papers cite this work, alongside 5 external citations. Polarity classification is still indexing.
19
Pith papers citing it
5
external citations · external index
hub tools
citation-role summary
background 1
citation-polarity summary
roles
background 1polarities
background 1representative citing papers
Turn-averaged SAEs reconstruct average activations over conversation turns to represent high-level turn characteristics with a fixed number of features, simplifying long-context interpretability compared to per-token SAEs.
TimeROME-DLM enables training-free knowledge editing in masked diffusion language models via temporal causal tracing and low-rank residual edit memory applied at inference time.
PRISM is a new activation-conditioned model that recovers full sets of simultaneous instructions from LLM hidden states via judge-guided GRPO training and outperforms prior activation-to-language methods on security-relevant tasks.
V-SEAM combines concept-level visual semantic editing with attention head modulation to identify positive and negative contributors across object, attribute, and relationship levels, then uses this to improve VLM performance on VQA benchmarks.
Authority sycophancy in LLMs is a layer-localized erasure of correct answer representations that scales with authority level and resists simple interventions.
Across four language models, the linear direction detecting hallucinated entities achieves perfect separability yet lies at cosine 0.12 from the direction that induces refusal, with the gap unchanged by instruction tuning.
Attribute retrieval in LLMs follows non-contiguous, redundant layer paths identified via iterative patching, implying highly distributed knowledge storage.
LLMs generate adequate counterspeech for co-occurring hate and misinformation in 40% of cases, with a mixed knowledge strategy from fact-checkers and NGOs proving most effective after expert revision.
A latent mediation framework with sparse autoencoders enables non-additive token-level influence attribution in LLMs by learning orthogonal features and back-propagating attributions.
Tool identity is linearly readable and steerable in LLMs via mean activation differences, with 77-100% switch accuracy and error prediction from activation gaps.
Knowledge conflicts in hypernetwork LLM adaptation stem from constant adapter margins losing to frequency-dependent pretrained margins; selective layer boosting and conflict-aware triggering raise deep-conflict accuracy to 71-72.5% on Gemma-2B and Mistral-7B.
LLM 2-bit quantization fails via either cumulative signal degradation or early computation collapse in key components.
Mechanistic experiments on Gemma 3 27B, Qwen 2.5 7B and Magistral Small 24B show verbal confidence is cached at post-answer positions from answer tokens and captures richer answer-quality information beyond token log-probabilities.
Balanced parametric and in-context knowledge use in LLMs is an emergent property requiring intra-document repetition, moderate inconsistency, and skewed distributions in training data.
Across 15 LLMs, first-answer accuracy on arithmetic procedural execution falls from 63% at 5 steps to 20% at 95 steps, with under-execution increasing.
Activation patching provides evidence about neural network circuits when the choice of metric is aligned with the hypothesis and common interpretation errors are avoided.
Varying evaluation metrics and corruption methods in activation patching produces different localization and circuit discovery outcomes in language models, leading to recommendations for preferred practices.
citing papers explorer
-
CATCH-ME if you RAG: a dataset of Contextually Annotated multi-Turn Counterspeech against Hate and Misinformation Exchanges
Presents a new expert-curated dataset of multi-turn counterspeech dialogues in five languages targeting hate against seven groups, with span annotations linking to verified external knowledge for RAG applications.
-
Turn-Averaged SAEs for Feature Discovery and Long-Context Attribution
Turn-averaged SAEs reconstruct average activations over conversation turns to represent high-level turn characteristics with a fixed number of features, simplifying long-context interpretability compared to per-token SAEs.
-
TimeROME-DLM: Temporal Causal Tracing and Low-Rank Inference-Time Knowledge Editing for Masked Diffusion Language Models
TimeROME-DLM enables training-free knowledge editing in masked diffusion language models via temporal causal tracing and low-rank residual edit memory applied at inference time.
-
PRISM: Recovering Instruction Sets from Language Model Activations
PRISM is a new activation-conditioned model that recovers full sets of simultaneous instructions from LLM hidden states via judge-guided GRPO training and outperforms prior activation-to-language methods on security-relevant tasks.
-
V-SEAM: Visual Semantic Editing and Attention Modulating for Causal Interpretability of Vision-Language Models
V-SEAM combines concept-level visual semantic editing with attention head modulation to identify positive and negative contributors across object, attribute, and relationship levels, then uses this to improve VLM performance on VQA benchmarks.
-
A Mechanistic View of Authority Hierarchy in LLM Sycophancy
Authority sycophancy in LLMs is a layer-localized erasure of correct answer representations that scales with authority level and resists simple interventions.
-
Perfect Detection, Failed Control: The Geometry of Knowing vs. Steering in Language Models
Across four language models, the linear direction detecting hallucinated entities achieves perfect separability yet lies at cosine 0.12 from the direction that induces refusal, with the gap unchanged by instruction tuning.
-
Factual Retrieval in LLMs Is a Redundant, Distributed and Non-Contiguous Process
Attribute retrieval in LLMs follows non-contiguous, redundant layer paths identified via iterative patching, implying highly distributed knowledge storage.
-
Assisted Counterspeech Writing at the Crossroads of Hate Speech and Misinformation
LLMs generate adequate counterspeech for co-occurring hate and misinformation in 40% of cases, with a mixed knowledge strategy from fact-checkers and NGOs proving most effective after expert revision.
-
Correcting Influence: Unboxing LLM Outputs with Orthogonal Latent Spaces
A latent mediation framework with sparse autoencoders enables non-additive token-level influence attribution in LLMs by learning orthogonal features and back-propagating attributions.
-
Tool Calling is Linearly Readable and Steerable in Language Models
Tool identity is linearly readable and steerable in LLMs via mean activation differences, with 77-100% switch accuracy and error prediction from activation gaps.
-
The Override Gap: A Magnitude Account of Knowledge Conflict Failure in Hypernetwork-Based Instant LLM Adaptation
Knowledge conflicts in hypernetwork LLM adaptation stem from constant adapter margins losing to frequency-dependent pretrained margins; selective layer boosting and conflict-aware triggering raise deep-conflict accuracy to 71-72.5% on Gemma-2B and Mistral-7B.
-
From Signal Degradation to Computation Collapse: Uncovering the Two Failure Modes of LLM Quantization
LLM 2-bit quantization fails via either cumulative signal degradation or early computation collapse in key components.
-
How do LLMs Compute Verbal Confidence
Mechanistic experiments on Gemma 3 27B, Qwen 2.5 7B and Magistral Small 24B show verbal confidence is cached at post-answer positions from answer tokens and captures richer answer-quality information beyond token log-probabilities.
-
How Training Data Shapes the Use of Parametric and In-Context Knowledge in Language Models
Balanced parametric and in-context knowledge use in LLMs is an emergent property requiring intra-document repetition, moderate inconsistency, and skewed distributions in training data.
-
When LLMs Stop Following Steps: A Diagnostic Study of Procedural Execution in Language Models
Across 15 LLMs, first-answer accuracy on arithmetic procedural execution falls from 63% at 5 steps to 20% at 95 steps, with under-execution increasing.
-
How to use and interpret activation patching
Activation patching provides evidence about neural network circuits when the choice of metric is aligned with the hypothesis and common interpretation errors are avoided.
-
Towards Best Practices of Activation Patching in Language Models: Metrics and Methods
Varying evaluation metrics and corruption methods in activation patching produces different localization and circuit discovery outcomes in language models, leading to recommendations for preferred practices.
- Not Just RLHF: Why Alignment Alone Won't Fix Multi-Agent Sycophancy