INDEX

Explanations

specific data structures and their representations in machine learning contexts

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

WebServlet

-0.63

 AssemblyCulture

-0.61

AddTagHelper

-0.59

 ujednoznacz

-0.58

 pleaſure

-0.57

 Italijanski

-0.57

 BorderSide

-0.56

verifyException

-0.56

GEBURTSDATUM

-0.54

 estekak

-0.54

POSITIVE LOGITS

dibles

0.36

ent

0.33

rendered

0.32

 flashed

0.32

antel

0.32

 plate

0.31

 Insel

0.31

plated

0.31

 forefront

0.31

Batch

0.31

Activations Density 0.049%

specific data structures and their representations in machine learning contexts

No Comments

No Known Activations