INDEX

Explanations

expressions indicating capability and method

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

verb + complements

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

#+#

-0.45

 AssemblyTitle

-0.41

Rhestr

-0.40

zyms

-0.39

 ब्रेकडाउन

-0.39

 nữa

-0.39

riment

-0.38

Abp

-0.38

heitsbild

-0.37

MIDDLEWARE

-0.37

POSITIVE LOGITS

 ویکی‌پدیا

0.48

 ungrateful

0.45

ftagPool

0.41

djangoproject

0.41

BSITE

0.40

 atheists

0.38

 Verpackung

0.38

ArgsConstructor

0.37

 exhibición

0.36

 econó

0.35

Activations Density 0.070%

expressions indicating capability and method

verb + complements

No Comments

No Known Activations