INDEX

Explanations

occurrences of specific locations and numerical data

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

coordinates and numbers

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

 prate

-0.45

ècie

-0.45

zzleHttp

-0.42

 magist

-0.42

layui

-0.41

 newArr

-0.41

 Iterable

-0.39

riezmann

-0.38

 chiffre

-0.38

 Hül

-0.38

POSITIVE LOGITS

httphttps

0.54

RegressionTest

0.53

findpost

0.52

ValueStyle

0.48

 emptying

0.47

 surla

0.47

Cross

0.46

المناصب

0.45

 soggior

0.45

enumi

0.44

Activations Density 0.000%

occurrences of specific locations and numerical data

coordinates and numbers

No Comments

No Known Activations