INDEX

Explanations

references to numeric values or quantities in various contexts

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

will help others

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

Negative Logits

principalColumn

-0.71

Viitteet

-0.65

setVerticalGroup

-0.56

 AssemblyProduct

-0.54

 виправивши

-0.52

XmlAccessorType

-0.52

 propOrder

-0.50

 pleaſure

-0.49

httphttps

-0.48

 ſame

-0.47

POSITIVE LOGITS

 EconPapers

0.40

 schu

0.40

accéder

0.36

 peindre

0.35

 مد

0.35

 finally

0.34

 Platt

0.34

__]

0.33

gum

0.32

 lastly

0.32

Activations Density 0.024%

references to numeric values or quantities in various contexts

will help others

No Known Activations