INDEX

Explanations

references to incidents and physical damage

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

impacts and destruction

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

 leçon

-0.41

 للمعارف

-0.41

 doty

-0.40

:✨

-0.38

InjectAttribute

-0.38

 teratur

-0.37

andidaten

-0.37

 Methuen

-0.37

 závis

-0.37

 frontière

-0.36

POSITIVE LOGITS

fastjson

0.52

 vPvB

0.44

posedge

0.44

 deſt

0.43

 nahilalakip

0.43

ことができました

0.42

<?

0.40

EDEFAULT

0.40

TextInputLayout

0.40

DataSnapshot

0.40

Activations Density 0.016%

references to incidents and physical damage

impacts and destruction

No Comments

No Known Activations