Invisible Spotlight

Agent Spotlight: Brian Onyango, a Senior Lead at Invisible

Brian Onyango

Senior Lead

Invisible Spotlight

"I’ve experienced massive professional growth."


Meet Brian, a 3 year veteran at Invisible based in Nairobi, and now a Senior Lead. We caught up with Brian to learn more.

Where are you located?

Nairobi, Kenya

What language(s) do you speak? 

English, Swahili, and Dholuo.

What did you do before coming to Invisible? 

I was a student. 

Why did you decide to join Invisible? 

I needed an income to pay my tuition fees!

What are Your Top 3 moments at Invisible?

01|

Getting an opportunity to lead a group of Agents for the first time in 2019.

02|

Being promoted to Senior lead in 2022.

03|

When Invisible Management visited Kenya to meet us in 2019.

How have you grown at Invisible when you compare your experience on Day 1 with your experience now?

When I joined on January 17th, 2019, I had no Google Sheets knowledge, and today I find it limiting! I’ve experienced massive professional growth.

How would you describe Invisible to someone thinking about joining the team?

Invisible is a group of change-makers and disruptors who are making the unthinkable possible.

What is it that keeps you at Invisible?

The passion for what I do and the Smiles on the faces of my clients!

Invisible is a group of change-makers and disruptors who are making the unthinkable possible.

We're so proud of your growth at Invisible, Brian! 

Are you interested in joining an elite team of global professionals?

Andrew Hull


Meet Brian, a 3 year veteran at Invisible based in Nairobi, and now a Senior Lead. We caught up with Brian to learn more.

Where are you located?

Nairobi, Kenya

What language(s) do you speak? 

English, Swahili, and Dholuo.

What did you do before coming to Invisible? 

I was a student. 

Why did you decide to join Invisible? 

I needed an income to pay my tuition fees!

Overview

LLM Task

Benchmark Dataset/Corpus

Sentiment Analysis

SST-1/SST-2

Natural Language Inference /  Recognizing Textual Entailment

Stanford Natural Language Inference Corpus (SNLI)

Named Entity Recognition

conll-2003

Question Answering

SQuAD

Machine Translation

WMT

Text Summarization

CNN/Daily Mail Dataset

Text Generation

WikiText

Paraphrasing

MRPC

Language Modelling

Penn Tree Bank

Bias Detection

StereoSet

LLM Task

Benchmark Dataset/Corpus

Common Metric

Dataset available at

Sentiment Analysis

SST-1/SST-2

Accuracy

https://huggingface
.co/datasets/sst2

Natural Language Inference /  Recognizing Textual Entailment

Stanford Natural Language Inference Corpus (SNLI)

Accuracy

https://nlp.stanford.edu
projects/snli/

Named Entity Recognition

conll-2003

F1 Score

https://huggingface.co/
datasets/conll2003

Question Answering

SQuAD

F1 Score, Exact Match, ROUGE

https://rajpurkar.github.i
o/SQuAD-explorer/

Machine Translation

WMT

BLEU, METEOR

https://machinetranslate
.org/wmt

Text Summarization

CNN/Daily Mail Dataset

ROUGE

https://www.tensorflow
.org/datasets/catalog/
cnn_dailymail

Text Generation

WikiText

BLEU, ROUGE

https://www.salesforce.
com/products/einstein/
ai-research/the-wikitext-dependency-language-modeling-dataset/

Paraphrasing

MRPC

ROUGE, BLEU

https://www.microsoft.
com/en-us/download/details.a
spx?id=52398

Language Modelling

Penn Tree Bank

Perplexity

https://zenodo.org/recor
d/3910021#.ZB3qdHbP
23A

Bias Detection

StereoSet

Bias Score, Differential Performance

https://huggingface.co/
datasets/stereoset

Table 1 - Example of some LLM tasks with common benchmark datasets and their respective metrics. Please note for many of these tasks, there are multiple benchmark datasets, some of which have not been mentioned here.

Metric Selection

Metric

Usage

Accuracy

Measures the proportion of correct predictions made by the model compared to the total number of predictions.

Precision

Measures the proportion of true positives out of all positive predictions.

Recall

Measures the proportion of true positives out of all actual positive instances.

F1 Score

Measures the harmonic mean of precision and recall.

Perplexity

Measures the model's uncertainty in predicting the next token (common in text generation tasks).

BLEU

Measures the similarity between machine-generated text and reference text.

ROUGE

Measures the similarity between machine-generated and human-generated text.

METEOR

May have higher computational complexity compared to BLEU or ROUGE.Requires linguistic resources for matching, which may not be available for all languages.

Pros

Cons

Simple interpretability. Provides an overall measure of model performance.

Sensitive to dataset imbalances, which can make it not informative. Does not take into account false positives and false negatives.

Useful when the cost of false positives is high. Measures the accuracy of positive predictions.

Does not take into account false negatives.Depends on other metrics to be informative (cannot be used alone).Sensitive to dataset imbalances.

Useful when the cost of false negatives is high.

Does not take into account false negatives.Depends on other metrics to be informative (cannot be used alone)and Sensitive to dataset imbalances.

Robust to imbalanced datasets.

Assumes equal importance of precision and recall.May not be suitable for multi-class classification problems with different class distributions.

Interpretable as it provides a single value for model performance.

May not directly correlate with human judgment.

Correlates well with human judgment.Easily interpretable for measuring translation quality.

Does not directly explain the performance on certain tasks (but correlates with human judgment).Lacks sensitivity to word order and semantic meaning.

Has multiple variants to capture different aspects of similarity.

May not capture semantic similarity beyond n-grams or LCS.Limited to measuring surface-level overlap.

Addresses some limitations of BLEU, such as recall and synonyms.

May have higher computational complexity compared to BLEU or ROUGE.Requires linguistic resources for matching, which may not be available for all languages.

Metric

Usage

Pros

Cons

Accuracy

Measures the proportion of correct predictions made by the model compared to the total number of predictions.

Simple interpretability. Provides an overall measure of model performance.

Sensitive to dataset imbalances, which can make it not informative. Does not take into account false positives and false negatives.

Precision

Measures the proportion of true positives out of all positive predictions.

Useful when the cost of false positives is high. Measures the accuracy of positive predictions.

Does not take into account false negatives.Depends on other metrics to be informative (cannot be used alone).Sensitive to dataset imbalances.

Recall

Measures the proportion of true positives out of all actual positive instances.

Useful when the cost of false negatives is high.

Does not take into account false negatives.Depends on other metrics to be informative (cannot be used alone)and Sensitive to dataset imbalances.

F1 Score

Measures the harmonic mean of precision and recall.

Robust to imbalanced datasets.

Assumes equal importance of precision and recall.May not be suitable for multi-class classification problems with different class distributions.

Perplexity

Measures the model's uncertainty in predicting the next token (common in text generation tasks).

Interpretable as it provides a single value for model performance.

May not directly correlate with human judgment.

BLEU

Measures the similarity between machine-generated text and reference text.

Correlates well with human judgment.Easily interpretable for measuring translation quality.

Does not directly explain the performance on certain tasks (but correlates with human judgment).Lacks sensitivity to word order and semantic meaning.

ROUGE

Measures the similarity between machine-generated and human-generated text.

Has multiple variants to capture different aspects of similarity.

May not capture semantic similarity beyond n-grams or LCS.Limited to measuring surface-level overlap.

METEOR

Measures the similarity between machine-generated translations and reference translations.

Addresses some limitations of BLEU, such as recall and synonyms.

May have higher computational complexity compared to BLEU or ROUGE.Requires linguistic resources for matching, which may not be available for all languages.

Table 2 - Common LLM metrics, their usage as a measurement tool, and their pros and cons. Note that for some of these metrics there exist different versions. For example, some of the versions of ROUGE include ROUGE-N, ROUGE-L, and ROUGE-W. For context, ROUGE-N measures the overlap of sequences of n-length-words between the text reference and the model-generated text. ROUGE-L measures the overlap between the longest common subsequence of tokens in the reference text and generated text, regardless of order. ROUGE-W on the other hand, assigns weights (relative importances) to longer common sub-sequences of common tokens (similar to ROUGE-L but with added weights). A combination of the most relevant variants of a metric, like ROUGE is selected for comprehensive evaluation.

Andrew Hull

Schedule a call to learn more about how Invisible might help your business grow while navigating uncertainty.

Schedule a Call
Request a Demo
Request a Demo
Request a Demo
Request a Demo
Request a Demo
Request a Demo
Request a Demo
Request a Demo
Request a Demo
Request a Demo
Request a Demo