INDEX

Explanations

elements related to instructions or organizational frameworks within technical or procedural contexts

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

development potential

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

RTEE

-0.42

RegistryLite

-0.39

 firebaseConfig

-0.37

FailureListener

-0.37

yled

-0.35

 TType

-0.35

 typelib

-0.35

DBNull

-0.34

cèse

-0.34

AnchorTagHelper

-0.34

POSITIVE LOGITS



0.55

 utafitiHapana

0.49

Personendaten

0.42

 Infórmanos

0.42

 Autorisations

0.41

httphttps

0.40

 EconPapers

0.40

 gevolg

0.40

avut

0.40

 disambiguazione

0.39

Activations Density 0.212%

elements related to instructions or organizational frameworks within technical or procedural contexts

development potential

No Comments

No Known Activations