INDEX

Explanations

negations or statements that indicate falsehood

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

not followed by adjective

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

 snippetHide

-0.42

EventQueue

-0.38

intios

-0.36

HasAnnotation

-0.35

hamshire

-0.34

 lenti

-0.34

kegaard

-0.34

BeginContext

-0.34

reportWebVitals

-0.34

unhofer

-0.33

POSITIVE LOGITS

 mergeFrom

0.52

 Extragalactic

0.46

jenost

0.46

usercontent

0.46

setVerticalGroup

0.45

Personensuche

0.45

 missione

0.44

 Infórmanos

0.44

 among

0.44

msgTypes

0.42

Activations Density 0.030%

negations or statements that indicate falsehood

not followed by adjective

No Comments

No Known Activations