INDEX

Explanations

instances of strong opinions or critique

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

negative evaluations

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

Personendaten

-0.62

 cận

-0.43

 yolu

-0.39

 opér

-0.39

 démoc

-0.38

 registration

-0.36

 uLocal

-0.36

 فريبيس

-0.35

 thanks

-0.35

 recognition

-0.34

POSITIVE LOGITS

 Worse

0.52

InitVars

0.51

 autorytatywna

0.50

Worse

0.49

 peggio

0.49

execSQL

0.49

FunctionFlags

0.49

 {}));

0.48

 worse

0.48

>&

0.48

Activations Density 0.021%

instances of strong opinions or critique

negative evaluations

No Comments

No Known Activations