INDEX

Explanations

references to detailed reports and extensive projects

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

detailed reports and projects

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

новништво

-0.38

 disambiguazione

-0.35

 geest

-0.35

indépendance

-0.34

rungsseite

-0.34

 jengibre

-0.34

 indépend

-0.33

 CreateTagHelper

-0.33

 indépendante

-0.33

 citoyens

-0.32

POSITIVE LOGITS

RTEX

0.56

 виправивши

0.55

argout

0.54

 complex

0.53

SequentialGroup

0.52

FormTagHelper

0.52

 للمعارف

0.52

IntoConstraints

0.51

 AttributeSet

0.51

Complex

0.50

Activations Density 0.089%

references to detailed reports and extensive projects

detailed reports and projects

No Comments

No Known Activations