stylistic change

MLLMs.html

@@ -193,7 +193,7 @@ <h1>Here's what we are going to covering:</h1>
 			<tr>
 				<td width=260px>
 
-					<b>Multimodal Large Language Models (MLLMs)</b> allow language-based reasoning over other modalities, which is predominantly images right now but will very soon include videos and even <a href="https://arxiv.org/abs/2309.16058" target="_blank">sensor data</a>. Recently, there has been a lot of interest in understanding the robustness of LLMs–the major backbone of multimodal LLMs–to variations in the input data – ranging from unintended changes (<a href="https://arxiv.org/abs/2310.11324" target="_blank">Sclar et al., 2023</a>) to adversarial ones (<a href="https://arxiv.org/abs/2307.15043" target="_blank">Zou et al., 2023</a>). In the same spirit, we conduct robustness evaluations on the zero-shot capabilities of multimodal LLMs to plausible changes in the input text.<br/>
+					<b>Multimodal Large Language Models (MLLMs)</b> allow language-based reasoning over other modalities, which is predominantly images right now but will very soon include videos and even <a href="https://arxiv.org/abs/2309.16058" target="_blank">sensor data</a>. Recently, there has been a lot of interest in understanding the robustness of LLMs–the major backbone of multimodal LLMs–to variations in the input data – ranging from unintended changes (<a href="https://arxiv.org/abs/2310.11324" target="_blank">Sclar et al., 2023</a>) to adversarial ones (<a href="https://arxiv.org/abs/2307.15043" target="_blank">Zou et al., 2023</a>). In the same spirit, we conduct robustness evaluations on the zero-shot capabilities of multimodal LLMs to plausible changes in the input text.<br/><br/>
 
 					<hr/>
 					<!-- <center> -->