INDEX

Explanations

concepts related to security, especially concerning the disabling of devices and privacy measures

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

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Taktlose

-0.58

IntoConstraints

-0.57

rungsseite

-0.56

AutoField

-0.55

RTEX

-0.54

 ModelExpression

-0.53

+#+

-0.48

IVEREF

-0.48

scriptcase

-0.47

TypedDataSet

-0.45

POSITIVE LOGITS

 disable

0.52

 disabling

0.48

 shielding

0.45

disable

0.44

 Disable

0.43

 DBNull

0.42

 disabled

0.41

 exemption

0.40

 exemptions

0.40

 counteract

0.39

Activations Density 0.249%

concepts related to security, especially concerning the disabling of devices and privacy measures

No Comments

No Known Activations