INDEX

Explanations

specific numeric values, quantities, and qualifications related to jobs or roles

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

technical descriptors and context

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

 ainfi

-0.63

 ſche

-0.60

 Roskov

-0.59

 Reſ

-0.57

 raiſ

-0.55

 Verſ

-0.55

afficheront

-0.54

 linkovi

-0.53

Transkript

-0.52

 ſtate

-0.52

POSITIVE LOGITS

tagHelperRunner

0.50

ViewImports

0.49

OGND

0.43

 كومونز

0.40

 arşivlendi

0.38

 للمعارف

0.36

 године

0.36

 Narrow

0.35

 narrow

0.35

 undivided

0.34

Activations Density 0.002%

specific numeric values, quantities, and qualifications related to jobs or roles

technical descriptors and context

No Comments

No Known Activations