INDEX

Explanations

references to technical documentation and support materials

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

finding evidence or details

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 9-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_9/width_131k/average_l0_121

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.9.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

 चीज़ों

-0.49

NUMX

-0.43

 فريبيس

-0.41

rdı

-0.39

kyo

-0.39

 skola

-0.39

transQ

-0.38

 Italijani

-0.37

jores

-0.36

 Stolz

-0.36

POSITIVE LOGITS

 find

0.75

 Find

0.64

find

0.63

 FIND

0.62

 finds

0.62

 encuentras

0.60

Find

0.60

 found

0.58

FIND

0.57

 finden

0.56

Activations Density 0.072%

references to technical documentation and support materials

finding evidence or details

No Comments

No Known Activations