INDEX

Explanations

mathematical expressions or symbols related to values and coordinates

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

(c) followed by )

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

 Marchand

-0.46

abcdefghijklmnop

-0.44

ungsver

-0.44

 ACCORD

-0.41

Bioaccumulative

-0.40

Bakgrunnsstoff

-0.40

Apo

-0.38

 Bons

-0.38

 Trip

-0.37

 Fellowship

-0.37

POSITIVE LOGITS

InjectAttribute

0.69

 nakalista

0.57

 EconPapers

0.55

HomeAsUpEnabled

0.52

 الحره

0.50

tagHelperRunner

0.47

tonode

0.47

PositiveButton

0.46

 alfombra

0.45

 ***!

0.45

Activations Density 0.020%

mathematical expressions or symbols related to values and coordinates

(c) followed by )

No Comments

No Known Activations