INDEX

Explanations

specific phrases and constructs related to emotional responses and personal reflections

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

tend to be

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

httphttps

-0.52

onViewCreated

-0.51



-0.48

findpost

-0.46

SharedCtor

-0.43

providedIn

-0.40

 iNdEx

-0.39

oneofs

-0.38

modelBuilder

-0.38

скажите

-0.38

POSITIVE LOGITS

 يتيمه

0.58

 مشين

0.41

CppMethod

0.41

󠁴

0.41

irsten

0.37

 mula

0.36

ributes

0.36

GNUC

0.35

цезда

0.35

 tamb

0.35

Activations Density 0.130%

specific phrases and constructs related to emotional responses and personal reflections

tend to be

No Comments

No Known Activations