INDEX

Explanations

references to roles, responsibilities, or actions related to tasks

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

photos online

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

 يتيمه

-0.42

 ویکی‌پدی

-0.36

лтамалар

-0.35

󠁴

-0.35

+#+#

-0.35

 utafitiHapana

-0.33

Selfer

-0.32

HandlerContext

-0.32

 roll

-0.32

/*---

-0.32

POSITIVE LOGITS

PerformLayout

0.60

RuleContext

0.55

ScopeManager

0.49

ParallelGroup

0.48

Gön

0.47

 zvuky

0.46

internalType

0.44

fromCharCode

0.43

انتهای

0.43

SequentialGroup

0.43

Activations Density 0.068%

references to roles, responsibilities, or actions related to tasks

photos online

No Comments

No Known Activations