INDEX

Explanations

common conjunctions and linking phrases in context

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

in, expression, target

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

elemField

-0.71

parsedMessage

-0.47

transQ

-0.47



-0.47

Parcelize

-0.46

 للمعارف

-0.45

 لينك

-0.43

WaitGroup

-0.41

 الرياضيه

-0.41

trooper

-0.40

POSITIVE LOGITS

NewUrlParser

0.46

 blinded

0.45

pushFollow

0.45

instruments

0.44

riften

0.43

itoare

0.43

OGND

0.41

ConfigureAwait

0.41

 Schles

0.41

instrument

0.39

Activations Density 0.267%

common conjunctions and linking phrases in context

in, expression, target

No Comments

No Known Activations