INDEX

Explanations

references to data robustness and performance comparisons in analysis or research contexts

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

controlled or permissive conditions

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

AndEndTag

-0.46

Facet

-0.44

oredCriteria

-0.44

Organ

-0.42

oltà

-0.42

morrow

-0.42

 صوتيه

-0.42

abilia

-0.42

Throw

-0.41

paghetti

-0.41

POSITIVE LOGITS

ArrowToggle

0.46

SourceChecksum

0.45

 esfor

0.42

 Acero

0.40

 tried

0.39

 pourtant

0.39

 تلاش

0.39

 supposed

0.38

 försö

0.38

yarnpkg

0.37

Activations Density 0.160%

references to data robustness and performance comparisons in analysis or research contexts

controlled or permissive conditions

No Comments

No Known Activations