INDEX

Explanations

terms related to technology and coding

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

specific acronyms and names

np_acts-logits-general · gemini-2.5-flash-lite

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 للمعارف

-0.88

webElementXpaths

-0.68

WebElementEntity

-0.65

 Infórmanos

-0.63

 AssemblyCulture

-0.63

 Taktlose

-0.62

 Мексичка

-0.60

 transfieras

-0.59

tagHelperRunner

-0.59

HasAnnotation

-0.56

POSITIVE LOGITS

 Fabrication

0.44

InMemory

0.43

 Ferreira

0.43

 Riesen

0.43

InfoList

0.42

Hal

0.42

bere

0.42

–>

0.42

parseLong

0.41

bergen

0.41

Activations Density 0.573%

terms related to technology and coding

specific acronyms and names

No Comments

No Known Activations