INDEX

Explanations

phrases and words that indicate temporal references or sequence of events

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

invention, comparison, or specific treatments

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

featureID

-0.88

 Италијани

-0.85

iſchen

-0.82

niſſe

-0.79

httphttps

-0.78

 ***!

-0.77

iſche

-0.77

 Wikimedijinoj

-0.77

WireFormatLite

-0.75

ſſung

-0.75

POSITIVE LOGITS

vœ

0.42

 désir

0.33

 adjunto

0.32

 derfor

0.32

 mož

0.32

 curieux

0.31

vábbi

0.31

 therefore

0.31

 Investigación

0.31

 Espíritu

0.31

Activations Density 0.129%

phrases and words that indicate temporal references or sequence of events

invention, comparison, or specific treatments

No Comments

No Known Activations