INDEX

Explanations

instances of conjunctions that signify connections between entities or concepts

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

conjunctions and disjunctions

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

gafas

-0.48

 Penga

-0.46

 burbujas

-0.46

 guadag

-0.45

 coisa

-0.45

 jemanden

-0.45

näm

-0.44

 Comunicação

-0.44

 cooperación

-0.43

wendigkeit

-0.43

POSITIVE LOGITS

webElementXpaths

0.82

httphttps

0.80

setVerticalGroup

0.75

StructEnd

0.64

LookAnd

0.63

 ligiloj

0.63

 ་་

0.61

__":

0.60

 للاسماء

0.60

Personendaten

0.59

Activations Density 0.110%

instances of conjunctions that signify connections between entities or concepts

conjunctions and disjunctions

No Comments

No Known Activations