INDEX

Explanations

references to data processing and resource management in a technical context

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

pool with numbers

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

 gdyby

-0.55

 vorige

-0.52

 fatos

-0.47

 Marten

-0.47

 zijne

-0.47

 veranderen

-0.47

 artigian

-0.46

 século

-0.46

 tijds

-0.45

 médié

-0.45

POSITIVE LOGITS

 pool

1.34

 Pool

1.25

Pool

1.23

pool

1.23

 POOL

1.20

POOL

1.07

 pools

1.06

池

1.01

 Pools

1.01

 pooling

0.99

Activations Density 0.008%

references to data processing and resource management in a technical context

pool with numbers

No Comments

No Known Activations