INDEX

Explanations

phrases that indicate relationships or connections between ideas

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

"and" followed by a negative word

np_acts-logits-general · gemini-2.0-flash

conjunctions and consequences

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-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 ſtand

-0.63

 ſou

-0.57

 ſeveral

-0.56

 disambiguazione

-0.55

 Houſe

-0.54

 pleaſure

-0.52

 Inscrivez

-0.52

 ſtre

-0.52

ſelf

-0.51

ſelves

-0.50

POSITIVE LOGITS

verifyException

0.42

utilisons

0.41

thâu

0.40

pyx

0.38

قایناقلار

0.36

 wanting

0.35

setViewportView

0.35

enx

0.35

 jeito

0.35

 autorytatywna

0.34

Activations Density 0.487%

phrases that indicate relationships or connections between ideas

"and" followed by a negative word

conjunctions and consequences

No Comments

No Known Activations