INDEX

Explanations

instances of self-reflection and personal expression in writing

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

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

游戏亮点

-0.36

 répon

-0.35

 Anſ

-0.35

 الحره

-0.34

 writ

-0.34

mögen

-0.33

 udrž

-0.33

はじめに

-0.33

 ciphertext

-0.32

 inversiones

-0.31

POSITIVE LOGITS

keyColumn

0.61

الحياه

0.56

 nakalista

0.53

 kaarangay

0.52

principalColumn

0.52

PerformLayout

0.50

 newOwner

0.48

AlterField

0.48

 afresh

0.48

 wireType

0.47

Activations Density 0.088%

instances of self-reflection and personal expression in writing

No Comments

No Known Activations