ML Research Wiki / Benchmarks / Interpretability Techniques for Deep Learning / CausalGym

CausalGym

Interpretability Techniques for Deep Learning Benchmark

Performance Over Time

📊 Showing 7 results | 📏 Metric: Log odds-ratio (pythia-6.9b)

Top Performing Models

Rank	Model	Paper	Log odds-ratio (pythia-6.9b)	Date	Code
1	DAS	CausalGym: Benchmarking causal interpretability methods on linguistic tasks	9.95	2024-02-19	📦 aryamanarora/causalgym 📦 jannik-brinkmann/multilingual-features
2	Linear probe	CausalGym: Benchmarking causal interpretability methods on linguistic tasks	3.42	2024-02-19	📦 aryamanarora/causalgym 📦 jannik-brinkmann/multilingual-features
3	Difference-in-means	CausalGym: Benchmarking causal interpretability methods on linguistic tasks	2.91	2024-02-19	📦 aryamanarora/causalgym 📦 jannik-brinkmann/multilingual-features
4	k-means	CausalGym: Benchmarking causal interpretability methods on linguistic tasks	1.87	2024-02-19	📦 aryamanarora/causalgym 📦 jannik-brinkmann/multilingual-features
5	PCA	CausalGym: Benchmarking causal interpretability methods on linguistic tasks	1.81	2024-02-19	📦 aryamanarora/causalgym 📦 jannik-brinkmann/multilingual-features
6	LDA	CausalGym: Benchmarking causal interpretability methods on linguistic tasks	0.27	2024-02-19	📦 aryamanarora/causalgym 📦 jannik-brinkmann/multilingual-features
7	Random	CausalGym: Benchmarking causal interpretability methods on linguistic tasks	0.01	2024-02-19	📦 aryamanarora/causalgym 📦 jannik-brinkmann/multilingual-features

All Papers (7)

CausalGym: Benchmarking causal interpretability methods on linguistic tasks

2024

DAS

aryamanarora/causalgym jannik-brinkmann/multilingual-features

CausalGym: Benchmarking causal interpretability methods on linguistic tasks

2024

Linear probe

aryamanarora/causalgym jannik-brinkmann/multilingual-features

CausalGym: Benchmarking causal interpretability methods on linguistic tasks

2024

Difference-in-means

aryamanarora/causalgym jannik-brinkmann/multilingual-features

CausalGym: Benchmarking causal interpretability methods on linguistic tasks

2024

k-means

aryamanarora/causalgym jannik-brinkmann/multilingual-features

CausalGym: Benchmarking causal interpretability methods on linguistic tasks

2024

PCA

aryamanarora/causalgym jannik-brinkmann/multilingual-features

CausalGym: Benchmarking causal interpretability methods on linguistic tasks

2024

LDA

aryamanarora/causalgym jannik-brinkmann/multilingual-features

CausalGym: Benchmarking causal interpretability methods on linguistic tasks

2024

Random

aryamanarora/causalgym jannik-brinkmann/multilingual-features

CausalGym

Performance Over Time

Edit Benchmark Results

Edit Result

Top Performing Models

All Papers (7)

CausalGym: Benchmarking causal interpretability methods on linguistic tasks

CausalGym: Benchmarking causal interpretability methods on linguistic tasks

CausalGym: Benchmarking causal interpretability methods on linguistic tasks

CausalGym: Benchmarking causal interpretability methods on linguistic tasks

CausalGym: Benchmarking causal interpretability methods on linguistic tasks

CausalGym: Benchmarking causal interpretability methods on linguistic tasks

CausalGym: Benchmarking causal interpretability methods on linguistic tasks

Model	Paper	Log odds-ratio (pythia-6.9b)	Date
DAS	CausalGym: Benchmarking causal interpretability m…	9.95	2024-02-19
Linear probe	CausalGym: Benchmarking causal interpretability m…	3.42	2024-02-19
Difference-in-means	CausalGym: Benchmarking causal interpretability m…	2.91	2024-02-19
k-means	CausalGym: Benchmarking causal interpretability m…	1.87	2024-02-19
PCA	CausalGym: Benchmarking causal interpretability m…	1.81	2024-02-19
LDA	CausalGym: Benchmarking causal interpretability m…	0.27	2024-02-19
Random	CausalGym: Benchmarking causal interpretability m…	0.01	2024-02-19