INDEX

Explanations

references to academic conferences and presentations

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

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

########.

-0.44

 препратки

-0.44

Дереккөздер

-0.42

 capace

-0.42

 camiset

-0.39

 Indah

-0.39

 alegria

-0.39

 conhece

-0.37

 memen

-0.36

 sonriente

-0.35

POSITIVE LOGITS

 cession

0.54

SBATCH

0.52

 conference

0.52

 Conference

0.51

 MainAxisSize

0.50

 onResponse

0.49

 MongoClient

0.49

 convention

0.49

AGM

0.48

tangentMode

0.47

Activations Density 0.016%

references to academic conferences and presentations

No Comments

No Known Activations