One of our developers' review "this conference was surprisingly good" - this is a HUGE compliment #dlfforum @clirnews.bsky.social
7
0
1
0
Hashtag
#DLFForum
Advertisement · 728 × 90
We're the last session so you could either get a jump on your trip to the airport or come to our session.
Lauren and I will make it worth your while 👿 #dlfforum
2
0
0
0
Not suitable for this task without lots of review and validation. Not a time saver. Low confidence from tool gave higher user confidence in answer. #DLFForum #DLF2025
1
0
0
0
Strengths - feature identification pretty good, general geographic identification ok. Opacity a real issue in reviewing results. Also verbose. Inconsistency on prompted task, and time consuming for people to review #DLFForum #DLF2025 At one point ChatGPT asked if it could do something else LOL
0
0
1
0
Prompt drift an issue with ChatGPT or would start referencing previous images. Prompt drift not an issue with other models #DLFForum #DLF2025 Aerial photos in sequence - models could recognize that but unable to judge the set as a group or course correct.
0
0
1
0
Claude would give broad answers that were more correct, other models would confidently mislabel images. #DLFForum #DLF2025
0
0
1
0
Chat CPT performed similarly for high context and low context prompt. Differences seen in other models. ChatGPT also stopped giving confidence scores #DLFForum #DLF2025
0
0
1
0
ChatGPT overall did the best, followed by Gemini, then Claude according to scoring system developed for the project. Confidence assigned by models - ChatGPT fairly high, Claude low. #DLFForum #DLF2025 - Claude confidence and correctness in alignment, ChatGPT and Gemini not in alignment.
1
0
1
0
Prompting issues - phrasing, context, specificity, and user role all important. Tried low context and high context prompts. Did not give extra guidance. #DLFForum #DLF2025
0
0
1
0
Fed LLM sequence of images and asked for geographic info like latitude and longitude coordinates, placenames. - mountain landscapes, aerial images - 20 samples of each #DLFForum #DLF2025
0
0
1
0
Used VPN to attempt to prevent LLM abilities to detect location by user query. Also created burner emails / new user accounts for all of these platforms. Not sure if this was effective, but an attempt was made!
#DLFForum #DLF2025
0
0
1
0
University of Colorado Boulder looking at three free models to see how they did with geographic resources like aerial maps. Wanted to use DeepSeek but it didn't do image analysis
#DLFForum #DLF2025
1
0
1
0
In session on Measuring and Critiquing genAI tools for Geographic Metadata Creation
#DLFForum #DLF2025
1
0
1
0
Responsible AI in Libraries and Archives team has released our interactive Viewfinder toolkit.
Toolkit (interactive website): www.lib.montana.edu/responsible-...
Project info: www.lib.montana.edu/responsible-...
#DLFForum #DLF2025
3
3
0
0
Reasoning with Small Language Models (SLM) to Create Trustworthy GenAI
Slides: doi.org/10.5281/zeno...
System Prompt: gist.github.com/jasonclark/1...
Model Context Protocol server:
gist.github.com/jasonclark/4...
#DLFForum #DLF2025
1
0
0
0
Important to keep context in mind when doing this type of assessment. What relationships to feature when describing a family photo. Example of inconsistent terms over time - farmer vs agricultural laborer with racial connotations. #DLFForum #DLF2025
0
0
0
0
Overview of process using lexicons and csv metadata with MaRMAT. Example of spreadsheet identifying aggrandizement terms like "pioneer" "gentleman" "famour" #DLFForum #DLF2025
0
0
1
0
Work on using MaRMAT for assessment of collections at University of Louisville. Note that @rjwcore.bsky.social has stickers!!! #DLFForum #DLF2025
2
0
1
0
Development of working group for harmful content statement at UNR had broad representation - 10 members across different departments, working on it for 6-8 months. #DLFForum #DLF2025 statement used to propel practical work such as work on Basque posters and local vocabulary
1
0
1
0
Harmful content statement elements - accessibility of statement to the public, scope of materials, acknowledgement of harm in a collection + metadata describing it, statement of future action/goals #DLFForum #DLF2025
0
0
1
0
Data Advocacy for All Toolkit da4all.github.io #DLFForum
3
0
1
0
Starting out by defining harmful, problematic, and outdated content - remember nuance between these definitions. #DLFForum #DLF2025
0
0
1
0
Hope to develop tool based on feedback and serving different use cases - community exists. Importance of social infrastructure for continued development of DH software. #DLFForum #DLF2025 Developers important but community also contributes use cases, feedback, etc. IIIF as a standard also key
0
0
1
0
How to work with spatial metadata at scale, how to deal with interoperability in IIF ecosystem, standardized spatial metadata. Termination of grant was 19 months into 3 year grant period. They are continuing to do work, but it is stalled without $$ for development #DLFForum #DLF2025
1
0
1
0
Mention of terminated NEH grant for development of Allmaps curator project :( - how to manage the annotations and ensure accuracy of the work.
#DLFForum #DLF2025
2
0
1
0
