INDEX

Explanations

contingency phrases and uncertainty in statements

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

technical or code identifiers

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

 fieltro

-0.40

 pasaje

-0.38

 devenus

-0.36

 principles

-0.36

 growing

-0.34

 commerciales

-0.32

 rescate

-0.32

 sürd

-0.32

 plantillas

-0.31

 creciendo

-0.31

POSITIVE LOGITS

 ویکی‌پدی

0.53

 ObjectParameter

0.51

AlterField

0.51

 мәкал

0.51

 AssemblyCulture

0.50

 <<<<<<<<<<<<<<

0.50

➋

0.48

 Normdatei

0.48

iastes

0.48

autaire

0.47

Activations Density 0.010%

contingency phrases and uncertainty in statements

technical or code identifiers

No Comments

No Known Activations