INDEX

Explanations

code-related syntax and structure, particularly focusing on function calls and method chaining

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

SQL keywords and commands

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

-0.49

DoubleQuotes

-0.45

 noDo

-0.43

 pleaſure

-0.40

awtextra

-0.40

 ExecuteAsync

-0.39

 ComVisible

-0.36

 ControllerBase

-0.36

SerializedName

-0.36

رشف

-0.36

POSITIVE LOGITS

 GenerationType

0.51

 survey

0.51

AnchorStyles

0.47

inSlope

0.47

udia

0.46

 selezion

0.46

 InputDecoration

0.45

XtraGrid

0.45

 Survey

0.44

 OPTIM

0.44

Activations Density 0.010%

code-related syntax and structure, particularly focusing on function calls and method chaining

SQL keywords and commands

No Comments

No Known Activations