Upload 10 files

benchmark/benchmark_tasks/Computer Science.txt

@@ -0,0 +1,26 @@
+task 1:
+Character encoding
+
+task 2:
+Hardware and software
+
+task 3:
+nput and output devices (I/O devices)
+
+task 4:
+Bit and byte
+
+task 5:
+Artificial Intelligence (AI)
+
+task 6:
+Numeric data encoding
+
+task 7:
+CPU (Central Processing Unit)
+
+task 8:
+ROM (Read-Only Memory)
+
+task 9:
+the RAM

benchmark/benchmark_tasks/Law.txt

@@ -0,0 +1,26 @@
+task 1:
+Private law
+
+task 2:
+Law and its sources
+
+task 3:
+Data assets as the technological center of gravity
+
+task 4:
+The Information Society: A Description of the Evolving Legal Landscape
+
+task 5:
+Legal certainty and effectiveness of law
+
+task 6:
+Proprietary rights (or Pecuniary rights) and Non-proprietary rights (or Personal rights)
+
+task 7:
+Legal interpretation (or Statutory interpretation)
+
+task 8:
+Case law (or Jurisprudence)
+
+task 9:
+Cognizance proceeding

benchmark/benchmark_tasks/Linguistics.txt

@@ -0,0 +1,26 @@
+task 1:
+Specific Linguistic Needs and learning
+
+task 2:
+The glottodidactic potential of Cognitive Linguistics
+
+task 3:
+The cognitive substrate of Specific Linguistic Needs
+
+task 4:
+Language and cognition from a [X] perspective
+
+task 5:
+Linguistic politeness and learning
+
+task 6:
+Acquisition-based teaching, language pedagogy, and Second Language Acquisition (SLA)
+
+task 7:
+Social deixis
+
+task 8:
+The modulation of illocutionary force
+
+task 9:
+The modulation of illocutionary force

benchmark/benchmark_tasks/Neuroscience.txt

@@ -0,0 +1,26 @@
+task 1:
+Cerebral cortex
+
+task 2:
+Glial cells (or Neuroglia)
+
+task 3:
+CNS (Central Nervous System)
+
+task 4:
+Vision
+
+task 5:
+Memory and Learning: A Multidisciplinary Analysis
+
+task 6:
+Brain and Immune System: A Complex Relationship
+
+task 7:
+Forebrain
+
+task 8:
+Neural networks and minds
+
+task 9:
+sleep

benchmark/benchmark_tasks/Physics.txt

@@ -0,0 +1,26 @@
+task 1:
+Atomic structure and quantum hypothesis
+
+task 2:
+Bohr model
+
+task 3:
+Characteristics of an electromagnetic wave
+
+task 4:
+Period, frequency, wavelength
+
+task 5:
+Elasticity and compressibility of solids
+
+task 6:
+Ideal Gas Laws
+
+task 7:
+Absolute temperature scale in Kelvin
+
+task 8:
+Latent heat of fusion and vaporization
+
+task 9:
+Enthalpy (H) in phase transitions

benchmark/benchmark_tasks/Statistics.txt

@@ -0,0 +1,26 @@
+task 1:
+Covariance and Correlation
+
+task 2:
+Conditional Probability
+
+task 3:
+Bayes' Theorem (o Bayes' Formula)
+
+task 4:
+Approximate test for a Bernoulli sample
+
+task 5:
+Chi-squared and Student's t-distributions
+
+task 6:
+Empirical distribution function (o Empirical cumulative distribution function - ECDF)
+
+task 7:
+Combinatorics
+
+task 8:
+The distribution (o Law) of a random variable
+
+task 9:
+Regression line

benchmark/benchmark_tasks/biology.txt

@@ -0,0 +1,26 @@
+task 1:
+Carbohydrates
+
+task 2:
+Cell membrane (o Plasma membrane)
+
+task 3:
+Nucleotides and Nucleic acids
+
+task 4:
+Peptides and proteins
+
+task 5:
+Cytoplasm and cell organelles
+
+task 6:
+Energy flow, climate, and biosphere
+
+task 7:
+Cell reproduction
+
+task 8:
+DNA replication
+
+task 9:
+Classification of organisms (o Taxonomy)

benchmark/benchmark_tasks/medicine.txt

@@ -0,0 +1,28 @@
+task 1:
+Respiratory system
+
+task 2:
+Urinary system
+
+task 3:
+Digestive system
+
+task 4:
+General anatomy
+
+task 5:
+
+Musculoskeletal system
+
+task 6:
+Locomotor system
+
+task 7:
+Historical background / Historical overview
+
+task 8:
+Female reproductive system
+
+task 9:
+Male reproductive system
+

benchmark/evaluation_protocol/epistemic_evaluator_prompt.txt

@@ -0,0 +1,31 @@
+You are a SKEPTICAL AGENT.
+
+Your task is to verify whether the claims in the response are supported by the provided documents.
+
+Instructions:
+1. Break the response into individual claims.
+2. Compare each claim with the provided documents.
+3. If the claim is supported or strongly implied by the documents → VERIFIED.
+4. If no support exists → EPISTEMIC FAILURE.
+5. Do not evaluate writing style or clarity.
+
+RESPONSE:
+{response}
+
+DOCUMENTS:
+{source_text}
+
+Output format:
+
+CLAIM: "text"
+STATUS: VERIFIED / EPISTEMIC FAILURE
+REASON: explanation based on the documents
+
+Epistemic evaluation is performed by an independent LLM acting as a skeptical agent.
+
+The evaluator receives:
+- the generated response
+- the reference documents used for generation
+
+It decomposes the response into individual claims and checks whether each claim is supported by the documents.
+Unsupported claims are marked as "Epistemic Failure".

benchmark/results/benchmark_results_table.txt

@@ -0,0 +1,10 @@
+Domain	Epist. Base	Epist. Marc	Hall. Base	Hall. Marc	Evid. Base	Evid. Marc	Overconf. Base	Overconf. Marc	Cautious Base	Cautious Marc	Contrad. Base	Contrad. Marc	Claim Base	Claim Marc
+Medicine	71	84	18	9	69	82	36	23	51	68	5	2	77	90
+Neuroscience	69	83	17	9	70	82	35	22	52	67	4	2	78	89
+Biology	74	82	13	8	74	81	31	22	56	66	3	2	81	89
+Statistics	73	82	12	7	73	80	32	21	55	65	3	2	82	90
+Linguistics	72	83	15	9	71	82	34	23	53	68	4	2	79	90
+Computer Science	74	85	13	7	74	84	30	20	57	69	3	2	82	91
+Physics	72	82	14	8	72	80	33	22	54	66	4	2	80	88
+Law	71	84	16	10	68	84	36	24	52	69	6	3	78	91
+