INDEX

Explanations

punctuation marks and formatting

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

historical periods or services

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

Spoljašnje

-0.45

Jeografia

-0.43

 Ghosh

-0.43

 Juli

-0.41

cur

-0.40

dou

-0.40

den

-0.40

 demon

-0.40

انتهای

-0.39

POSITIVE LOGITS

 antigüedad

0.57

 avoient

0.54

AccessorTable

0.50

 supérieurs

0.50

 actualidad

0.46

 pouvoit

0.46

 feroit

0.45

 CreateTagHelper

0.45

 sekitarnya

0.44

 EconPapers

0.44

Activations Density 0.058%

punctuation marks and formatting

historical periods or services

No Comments

No Known Activations