INDEX

Explanations

phrases related to purpose and reasons within a context

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

for which parties

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

 ModelExpression

-0.62

ScopeManager

-0.52

UnusedPrivate

-0.51

AndEndTag

-0.48

msgTypes

-0.46

OGND

-0.46

ilter

-0.46

SequentialGroup

-0.45

 שוליים

-0.45

iertamente

-0.45

POSITIVE LOGITS

 visit

0.42

Pourquoi

0.36

Vidite

0.35

曖昧さ回避

0.35

//};

0.35

 hitch

0.34

 doctor

0.34

'}}>

0.34

 haba

0.34

Solución

0.34

Activations Density 0.053%

phrases related to purpose and reasons within a context

for which parties

No Comments

No Known Activations