INDEX

Explanations

causal relationships and implications in arguments or statements

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

inadequately protected

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

 ModelExpression

-0.52

 للاسماء

-0.51

 ProtoMessage

-0.49

featureID

-0.46

LookAnd

-0.45

abestanden

-0.40

timedelta

-0.40

 Walkover

-0.40

 HttpNotFound

-0.38

InsertCommand

-0.38

POSITIVE LOGITS

 mauva

0.49

 deſt

0.46

 insegu

0.41

 évêque

0.41

цтва

0.40

archiviato

0.38

 lebaran

0.37

 worsened

0.37

 vícti

0.36

Spéc

0.36

Activations Density 0.160%

causal relationships and implications in arguments or statements

inadequately protected

No Comments

No Known Activations