INDEX

Explanations

terms related to lossless methods in various contexts

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

words ending in less

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 9-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_9/width_131k/average_l0_121

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.9.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

extAlignment

-0.54

 IActionResult

-0.53

 étoient

-0.53

 roomId

-0.50

bibfnamefont

-0.50

 MenuItem

-0.50

tranno

-0.49

 NgModule

-0.49

Agenda

-0.49

 Agenda

-0.48

POSITIVE LOGITS

less

2.86

LESS

2.14

Less

1.75

 Less

1.59

 less

1.52

lessness

1.45

 LESS

1.44

ELESS

1.34

eless

1.27

lessly

1.25

Activations Density 0.007%

terms related to lossless methods in various contexts

words ending in less

No Known Activations