INDEX

Explanations

specific articles and prepositions indicating common subjects or actions

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

in the

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

 conformidad

-0.39

 begitu

-0.38

 lendemain

-0.37

 antaranya

-0.37

 faudrait

-0.36

 antemano

-0.36

 tournage

-0.36

antaranya

-0.35

 Komunikasi

-0.35

accompagnement

-0.35

POSITIVE LOGITS

⟬

0.65

脚注の使い方

0.62

 パンチラ

0.60

<unused79>

0.59

<unused8>

0.59

[@BOS@]

0.59

<unused41>

0.59

<unused42>

0.59

<unused28>

0.59

<unused14>

0.59

Activations Density 0.122%

specific articles and prepositions indicating common subjects or actions

in the

No Comments

No Known Activations