INDEX

Explanations

terms related to specific numerical and structural relationships

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

visits / visiting

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

AutoScaleMode

-0.56

 obé

-0.55

Geplaatst

-0.55

 ainfi

-0.53

 myſelf

-0.52

 houſe

-0.52

awtextra

-0.50

parsedMessage

-0.50

leſs

-0.49

 propOrder

-0.49

POSITIVE LOGITS

httphttps

0.47

 visitó

0.42

hastly

0.35

ajevo

0.32

 visita

0.32

kfree

0.31

 visited

0.31

 visit

0.29

 visits

0.28

 visiting

0.28

Activations Density 0.221%

terms related to specific numerical and structural relationships

visits / visiting

No Comments

No Known Activations