INDEX

Explanations

actions and processes related to the assembly and positioning of components in a technical context

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

placed into or onto

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

ьаж

-0.66

Xna

-0.58

fontName

-0.53

 étoient

-0.52

 يتيمه

-0.52

 accountId

-0.52

jfree

-0.51

ftant

-0.50

 avoient

-0.50

عرِّف

-0.49

POSITIVE LOGITS

complexContent

0.43

StartTag

0.35

 surla

0.32

run

0.30

webElementXpaths

0.29

évaluateur

0.28

拨

0.28

run

0.27

 giving

0.27

 halfway

0.27

Activations Density 0.049%

actions and processes related to the assembly and positioning of components in a technical context

placed into or onto

No Comments

No Known Activations