INDEX

Explanations

instances of the word "and"

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

pronoun and pronoun

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

 pesquisa

-0.43

 surla

-0.41

IActionResult

-0.41

 resolution

-0.40

 investigation

-0.40

 regionales

-0.39

ValueGeneration

-0.38

 investigação

-0.37

 rate

-0.37

 rápida

-0.36

POSITIVE LOGITS

them

0.60

Him

0.59

DockStyle

0.58

principalTable

0.56

 Alfonso

0.54

Them

0.53

icar

0.52

olymp

0.51

Ben

0.50

your

0.50

Activations Density 0.025%

instances of the word "and"

pronoun and pronoun

No Comments

No Known Activations