INDEX

Explanations

terms related to education and artistic practices

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

learning and instruction

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

NUMX

-0.66

 nonUne

-0.59

𑄮

-0.58

脚注の使い方

-0.58

 CallOverrides

-0.58

𑄧

-0.57

𑄴

-0.57

𑄢

-0.57

afficheront

-0.57



-0.57

POSITIVE LOGITS

 skills

0.46

 training

0.46

 kurs

0.40

 instruction

0.40

 taught

0.40

 Training

0.40

 teaching

0.40

 cursos

0.38

 improvement

0.38

training

0.38

Activations Density 0.022%

terms related to education and artistic practices

learning and instruction

No Comments

No Known Activations