INDEX

Explanations

specific references or examples within a context, particularly focused on indications, instructions, or comparisons

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

following snippet

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

 Administrativna

-0.56

 يتيمه

-0.54

BeginContext

-0.51

 للاسماء

-0.50

VersionUID

-0.49

 مرئيه

-0.48

RegressionTest

-0.48

 الاطلاع

-0.47

Билгалдахарш

-0.46

RTLU

-0.45

POSITIVE LOGITS

thâu

0.36

enschappen

0.36

ţă

0.36

ParallelGroup

0.36

 grosse

0.35

myArray

0.35

ardige

0.35

avice

0.35

Honestly

0.34

lgica

0.34

Activations Density 0.102%

specific references or examples within a context, particularly focused on indications, instructions, or comparisons

following snippet

No Comments

No Known Activations

specific references or examples within a context, particularly focused on indications, instructions, or comparisons

following snippet

No Comments

No Known Activations