INDEX

Explanations

programming structures, such as functions and method definitions

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

function calls and returns

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

HandleFunc

-0.29

 Bereits

-0.27

ंदीखरीदारी

-0.25

 seç

-0.24

fjspx

-0.24

 otoño

-0.24

 placement

-0.23

 vẫn

-0.23

 hjem

-0.23

UrlResolution

-0.23

POSITIVE LOGITS

$_(

0.83

ðsíða

0.81

 kasarigan

0.77

iſche

0.74

iſchen

0.74

ſelben

0.74

$_=

0.74

ſicht

0.74

 ddelwed

0.73

 dieſes

0.73

Activations Density 0.020%

programming structures, such as functions and method definitions

function calls and returns

No Comments

No Known Activations