INDEX

Explanations

code snippets and programming-related dialogue

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

python REPL `>>>` and `isinstance`

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

 pleaſure

-0.58

 Majefty

-0.54

 ſta

-0.53

 fhort

-0.47

 kasarigan

-0.47

 houſe

-0.47

 ſtre

-0.47

 Houſe

-0.45

 itſelf

-0.45

 intStringLen

-0.44

POSITIVE LOGITS

 Gegen

0.50

 thử

0.48

LookAnd

0.47

ลอง

0.46

ISupport

0.44

 Upon

0.43

 ब्रेकडाउन

0.43

combine

0.41

 examined

0.41

对比

0.41

Activations Density 0.212%

code snippets and programming-related dialogue

python REPL `>>>` and `isinstance`

No Comments

No Known Activations