INDEX

Explanations

quotations and rhetorical questions in the text

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

past tense or predictions

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

IUrlHelper

-0.44

licó

-0.42

Хьажоргаш

-0.42

 Wiktionnaire

-0.41

 cerimonia

-0.41

expandindo

-0.40

 Butte

-0.40

Чыганаклар

-0.39

oció

-0.38

 diensten

-0.38

POSITIVE LOGITS

MigrationBuilder

0.71

 useAppContext

0.47

__(/*!

0.47

 Italijanski

0.40

 optimised

0.40

AddTagHelper

0.39

 AssemblyTitle

0.38

HandlerContext

0.38

Scénario

0.37

 privatisation

0.37

Activations Density 0.002%

quotations and rhetorical questions in the text

past tense or predictions

No Comments

No Known Activations