slides · global inventory
Diagnose
Slide function collection for the active corpus filters, with each row opening the slide detail view.
Matching slides 2,410
Current window 1,981-2,010
Pages 81
Rows per page 30
Slide function
1,981-2,010 of 2,410
47
The slide uses data from LAMA and SuperGlue benchmarks to illustrate performance variability in prompting.2021 Air Street Capital The State of AI Report 2021 · Air Street Capital
Consulting deck VC research Beat · The Facts (What)Loop · Pattern Hunter1 hit
51
The slide highlights the unreliability of human benchmarks in NLP, specifically referencing GPT-2 and GPT-3 evaluation.2021 Air Street Capital The State of AI Report 2021 · Air Street Capital
Consulting deck VC research Beat · The Facts (What)Loop · Pattern Hunter1 hit
53
The slide uses two charts to illustrate demographic imbalances: a population pyramid by ethnicity and age, and a stacked bar chart showing sex distribution by ethnicity.2021 Air Street Capital The State of AI Report 2021 · Air Street Capital
Consulting deck VC research Beat · The Facts (What)Loop · Iceberg1 hit
54
The slide uses a scatter plot to compare sensitivity vs false positive rate across various studies.2021 Air Street Capital The State of AI Report 2021 · Air Street Capital
Consulting deck VC research Beat · The Facts (What)Loop · Iceberg1 hit
58
The slide uses a diagram to propose a documentation framework for CommonCrawl-based datasets.2021 Air Street Capital The State of AI Report 2021 · Air Street Capital
Consulting deck VC research Beat · The Facts (What)Loop · Iceberg1 hit
67
The chart shows a bubble plot comparing ROC-AUC score vs GPU Memory (GB) for various GNN models.2021 Air Street Capital The State of AI Report 2021 · Air Street Capital
Consulting deck VC research Beat · The Facts (What)Loop · Pattern Hunter1 hit
109
The slide references a specific research paper/finding regarding underspecification in ML models.2021 Air Street Capital The State of AI Report 2021 · Air Street Capital
Consulting deck VC research Beat · The Facts (What)Loop · Iceberg1 hit
139
Includes a diagram illustrating the cohort assignment process.2021 Air Street Capital The State of AI Report 2021 · Air Street Capital
Consulting deck VC research Beat · The Facts (What)1 hit
20
The slide references a Princeton study on data leakage in 329 papers across 17 fields.2022 Air Street Capital The State of AI Report 2022 · Air Street Capital
Consulting deck VC research Beat · The Facts (What)Loop · Pattern Hunter1 hit
104
The slide uses the CoinRun experiment to illustrate a specific AI safety failure mode.2022 Air Street Capital The State of AI Report 2022 · Air Street Capital
Consulting deck VC research Beat · The Implications (So What)Loop · Pattern Hunter1 hit
22
The slide presents a critique of the 'emergent capabilities' phenomenon in LLMs, citing Stanford research that suggests these are artifacts of metric choice.2023 Air Street Capital The State of AI Report 2023 · Air Street Capital
Consulting deck VC research Beat · The Facts (What)Loop · Zoom In1 hit
24
The slide references the 'Lost in the Middle' research paper by Samaya.ai, UC Berkeley, Stanford, and LMSYS.org.2023 Air Street Capital The State of AI Report 2023 · Air Street Capital
Consulting deck VC research Beat · The Facts (What)Loop · Zoom In1 hit
147
The slide uses a specific case study (Sydney/Bing) to illustrate broader safety concerns like jailbreaking and prompt injection.2023 Air Street Capital The State of AI Report 2023 · Air Street Capital
Consulting deck VC research Beat · The Implications (So What)Loop · Pattern Hunter1 hit
149
The slide categorizes challenges into three pillars based on RLHF components, referencing specific sections (§3.1, §3.2, §3.3) of a research paper.2023 Air Street Capital The State of AI Report 2023 · Air Street Capital
Consulting deck VC research Beat · The Implications (So What)Loop · Pattern Hunter1 hit
152
The slide uses a complexity-theory-inspired notation (NP^NP) to illustrate the recursive difficulty of evaluating AI assistants.2023 Air Street Capital The State of AI Report 2023 · Air Street Capital
Consulting deck VC research Beat · The Implications (So What)Loop · Pattern Hunter1 hit
154
The chart shows a negative correlation between subject model parameters and explanation score.2023 Air Street Capital The State of AI Report 2023 · Air Street Capital
Consulting deck VC research Beat · The Implications (So What)Loop · Pattern Hunter1 hit
18
The slide uses two charts to demonstrate the impact of contamination: a bar chart showing performance gaps between standard benchmarks and clean test sets, and a scatter plot comparing exam scores to benchmark performance.2024 Air Street Capital The State of AI Report 2024 · Air Street Capital
Consulting deck VC research Beat · The Facts (What)Loop · Pattern Hunter1 hit
19
Includes a small visual grid of examples showing 'Bad Question Clarity', 'Bad Options Clarity', 'No Correct Answer', and 'Wrong Groundtruth'.2024 Air Street Capital The State of AI Report 2024 · Air Street Capital
Consulting deck VC research Beat · The Facts (What)Loop · Pattern Hunter1 hit
31
Includes a diagram illustrating the feedback loop of model collapse.2024 Air Street Capital The State of AI Report 2024 · Air Street Capital
Consulting deck VC research Beat · The Facts (What)Loop · Pattern Hunter1 hit
35
The slide includes a schematic diagram of the Ragnarök framework and a list of example research questions.2024 Air Street Capital The State of AI Report 2024 · Air Street Capital
Consulting deck VC research Beat · The Facts (What)Loop · Pattern Hunter1 hit
186
Includes a diagram illustrating a 'sleeper agent' attack mechanism.2024 Air Street Capital The State of AI Report 2024 · Air Street Capital
Consulting deck VC research Beat · The Implications (So What)Loop · Pattern Hunter1 hit
187
Includes a technical diagram (Figure 1) illustrating the transformation of model outputs into accuracy metrics.2024 Air Street Capital The State of AI Report 2024 · Air Street Capital
Consulting deck VC research Beat · The Implications (So What)Loop · Pattern Hunter1 hit
189
Includes a concrete example of a ChatGPT-4 interaction demonstrating sycophancy.2024 Air Street Capital The State of AI Report 2024 · Air Street Capital
Consulting deck VC research Beat · The Implications (So What)Loop · Pattern Hunter1 hit
190
The slide uses a series of scatter plots with fitted curves to demonstrate the relationship between KL divergence and win rates, suggesting that DPO is susceptible to over-optimization.2024 Air Street Capital The State of AI Report 2024 · Air Street Capital
Consulting deck VC research Beat · The Implications (So What)Loop · Pattern Hunter1 hit
193
The slide references specific research methodologies (semantic entropy and Google DeepMind's SAFE) to diagnose and mitigate LLM reliability issues.2024 Air Street Capital The State of AI Report 2024 · Air Street Capital
Consulting deck VC research Beat · The Implications (So What)Loop · Pattern Hunter1 hit
195
The slide discusses uncertainty calibration in LLMs using fine-tuning techniques like LoRA.2024 Air Street Capital The State of AI Report 2024 · Air Street Capital
Consulting deck VC research Beat · The Implications (So What)Loop · Pattern Hunter1 hit
197
The slide references a specific experiment with 45 occurrences out of 32,768 trials.2024 Air Street Capital The State of AI Report 2024 · Air Street Capital
Consulting deck VC research Beat · The Implications (So What)Loop · Pattern Hunter1 hit
201
Discusses the dual-use nature of interpretability research in AI safety.2024 Air Street Capital The State of AI Report 2024 · Air Street Capital
Consulting deck VC research Beat · The Implications (So What)Loop · Pattern Hunter1 hit
203
The slide highlights the dual-use nature of AI tools in biology (e.g., RFDiffusion) and the emerging governance response.2024 Air Street Capital The State of AI Report 2024 · Air Street Capital
Consulting deck VC research Beat · The Implications (So What)Loop · Pattern Hunter1 hit
20
The chart shows 'Reported vs Measured Results' and 'Variance DS-R1-1.5B' to demonstrate that performance swings are often due to variance rather than actual model improvement.2025 Air Street Capital The State of AI Report 2025 · Air Street Capital
Consulting deck VC research Beat · The Facts (What)Loop · Zoom In1 hit