INDEX

Explanations

conditional phrases and instructions for behavior

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

if conditionals

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

verwijspagina

-0.47

HandleFunc

-0.46

 recensement

-0.38

 citoy

-0.36

 éditoriale

-0.35

}`).

-0.35

 distinta

-0.34

 tercero

-0.33

ScopeManager

-0.33

 prises

-0.32

POSITIVE LOGITS

Portail

0.60

KURZBESCHREIBUNG

0.59

BagLayout

0.58

ArgumentParser

0.57

CrossRef

0.53

0.52

TAINMENT

0.51

[]:

0.51

脚注の使い方

0.51

<?

0.49

Activations Density 0.015%

conditional phrases and instructions for behavior

if conditionals

No Comments

No Known Activations