INDEX

Explanations

programming-related terms and constructs

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

static final constants

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

ográficas

-0.48

eachtet

-0.46

Autoritní

-0.46

 ſta

-0.45

principalColumn

-0.44

ApplicationTests

-0.44

 PeEnEo

-0.44

 Meksiku

-0.43

afbeelding

-0.42

clerView

-0.42

POSITIVE LOGITS

0.44

rahat

0.39

Handlung

0.38

achelor

0.37

 المعيارى

0.36

StructEnd

0.36

هرست

0.36

Opere

0.35

 cherchés

0.35

 rzecz

0.35

Activations Density 0.012%

programming-related terms and constructs

static final constants

No Comments

No Known Activations