INDEX

Explanations

patterns and structures related to data organization or representation

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

pair of roots

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

posedge

-0.66

]+$

-0.54

uxxxx

-0.53

 jSON

-0.53

parsedMessage

-0.48

adpleegd

-0.47

 infan

-0.46

=$

-0.45

))/(

-0.45

 esternos

-0.45

POSITIVE LOGITS

0.54

 ambientale

0.50

 padat

0.49

 стоковые

0.49

 mutiara

0.49

 tecnici

0.47

}{|

0.47

śred

0.46

 bedo

0.46

 dourada

0.45

Activations Density 0.003%

patterns and structures related to data organization or representation

pair of roots

No Comments

No Known Activations