INDEX

Explanations

references to repetition or recurrence in various contexts

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

again

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

 Guid

-0.37

 EconPapers

-0.36

стоин

-0.35

estris

-0.34

Sein

-0.34

ogenen

-0.34

 económico

-0.34

 Dari

-0.34

出版年

-0.33

 média

-0.33

POSITIVE LOGITS

RenderAtEndOf

0.61

 Италијани

0.51

msgTypes

0.48

олові

0.47

 kaarangay

0.47

 wikipagina

0.46

XMLSchema

0.45



0.45

 вікісторінку

0.44

creva

0.44

Activations Density 0.035%

references to repetition or recurrence in various contexts

again

No Comments

No Known Activations