INDEX

Explanations

technical terms related to error correction coding

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

code dimension and decoders

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

Not in Any Lists

No Comments

Negative Logits

 esternos

-0.45

 judiciales

-0.39

ształ

-0.38

oa̍t

-0.38

DotNetBar

-0.35

今

-0.34

optionalTypeArgs

-0.33

ppelin

-0.33

 academic

-0.32

BeforeClass

-0.32

POSITIVE LOGITS

bootstrapcdn

0.48

richTextPanel

0.45

Cordialement

0.44

Демографія

0.41

Heck

0.40

 itemprop

0.40

 CommonModule

0.40

 propOrder

0.39

 Heavens

0.39

sej

0.39

Activations Density 0.175%

technical terms related to error correction coding

code dimension and decoders

No Comments

No Known Activations