INDEX

Explanations

code and programming-related elements

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

javascript functions and user input

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

 nahilalakip

-0.61

 ProtoMessage

-0.59

 queſta

-0.54

TokenNameLBRACE

-0.53

 saites

-0.53

extAlignment

-0.53

tslib

-0.51

BeginContext

-0.51

tagHelperRunner

-0.50

 otomatig

-0.50

POSITIVE LOGITS

OnClickListener

0.37

 Fire

0.36

Once

0.34

σης

0.34

aler

0.34

Fire

0.33

tie

0.33

外

0.32

 again

0.32

 Operation

0.32

Activations Density 0.005%

code and programming-related elements

javascript functions and user input

No Comments

No Known Activations