6 papers
CANARY: Zero-Label Detection of Fine-Tuning Contamination in Language Models
Swapnil Parekh
Adversaries can implant latent harmful behavior by poisoning as few as 1% of fine-tuning examples. The contamination is invisible to every output-level defense: harmful behavior li…
Drop the Act: Probe-Filtered RL for Faithful Chain-of-Thought Reasoning
Swapnil Parekh
Reasoning models post-hoc rationalize answers they have already committed to internally, producing chains of *reasoning theater*: deliberative-looking steps that contribute nothing…
Thinking Wrong in Silence: Backdoor Attacks on Continuous Latent Reasoning
Swapnil Parekh
A new generation of language models reasons entirely in continuous hidden states, producing no tokens and leaving no audit trail. We show that this silence creates a fundamentally…
CIRCUS: Circuit Consensus under Uncertainty via Stability Ensembles
Swapnil Parekh
Every mechanistic circuit carries an invisible asterisk: it reflects not just the model's computation, but the analyst's choice of pruning threshold. Change that choice and the cir…
ACES: Accent Subspaces for Coupling, Explanations, and Stress-Testing in Automatic Speech Recognition
Swapnil Parekh
ASR systems exhibit persistent performance disparities across accents, but whether these gaps reflect superficial biases or deep structural vulnerabilities remains unclear. We intr…
CaptionFool: Universal Image Captioning Model Attacks
Swapnil Parekh
Image captioning models are encoder-decoder architectures trained on large-scale image-text datasets, making them susceptible to adversarial attacks. We present CaptionFool, a nove…