INDEX

Explanations

references to JSON and protocol buffers in coding contexts

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

json, proto, and ng syntax

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

 betweenstory

-1.17

 queſta

-0.84

transQ

-0.67

Personensuche

-0.65

 WebDriverWait

-0.62

########.

-0.60

 EconPapers

-0.60

lapsingToolbar

-0.60

 ویکی‌پدی

-0.59

 CanadaChoose

-0.59

POSITIVE LOGITS

JsonProperty

0.55

ApiModelProperty

0.53

 nahilalakip

0.42

 merak

0.40

 selten

0.39

 embuti

0.39

impanan

0.37

ع

0.36

HasColumnName

0.35

visible

0.35

Activations Density 0.001%

references to JSON and protocol buffers in coding contexts

json, proto, and ng syntax

No Comments

No Known Activations