INDEX

Explanations

structured code constructs, particularly those related to function definitions and control flow

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

code snippets

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

 defaultstate

-0.91

GOTREF

-0.82

rungsseite

-0.78

UrlResolution

-0.73

 kaarangay

-0.73

ⓧ

-0.68

Geplaatst

-0.66

 Chwiliwch

-0.66

 surla

-0.65

 Waſſer

-0.63

POSITIVE LOGITS

 fire

0.29

 https

0.28

 meant

0.28

 exactly

0.27

 himself

0.27

🖨

0.27

 precisely

0.26

це

0.26

by

0.25

 Mark

0.25

Activations Density 0.004%

structured code constructs, particularly those related to function definitions and control flow

code snippets

No Comments

No Known Activations