INDEX

Explanations

references to "the."

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

what's the

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

postIndex

-0.48

Rüyada

-0.48

fromnode

-0.46

pagest

-0.43

posedge

-0.39

parsedMessage

-0.39

thâu

-0.38

だけではなく

-0.37

Technologie

-0.37

erdings

-0.37

POSITIVE LOGITS

 maximum

0.59

the

0.58

 års

0.56

 expectations

0.56

 consequences

0.55

 Maximum

0.55

 requirements

0.54

有什么

0.54

ìm

0.54

Apakah

0.53

Activations Density 0.028%

references to "the."

what's the

No Comments

No Known Activations