INDEX

Explanations

phrases that emphasize familiarity and knowledge in a professional context

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

some type, into specific

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

Whilst

-0.50

 downwards

-0.47

 amongst

-0.45

parseFrom

-0.44

 Whilst

-0.44

 onwards

-0.43

 firstly

-0.43

 inception

-0.43

 monotonous

-0.42

 whilst

-0.41

POSITIVE LOGITS

 kasarigan

0.61

 Италијани

0.59

RTLR

0.47

RegressionTest

0.46

BagLayout

0.45

oredCriteria

0.44

ftagPool

0.44

újo

0.43

styleType

0.42

 retraite

0.42

Activations Density 0.030%

phrases that emphasize familiarity and knowledge in a professional context

some type, into specific

No Comments

No Known Activations