INDEX

Explanations

references to entities, roles, and attributes related to membership and identity

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

following the articles "a", "an", and sometimes "the"

np_acts-logits-general · gemini-2.0-flash

articles followed by specific words

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.18.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

 effort

-0.42

来自

-0.41

Ever

-0.40

UnusedPrivate

-0.40

 Ever

-0.40

 formerly

-0.40

 القدر

-0.39

Ability

-0.39

ance

-0.39

ANCE

-0.39

POSITIVE LOGITS

Datuak

1.12

 ویکی‌پدیا

0.92

 disambiguazione

0.89

tvguidetime

0.85

لينكات

0.81

 CreateTagHelper

0.80

StoreMessageInfo

0.80

تقاوى

0.78

 EconPapers

0.77

 виправивши

0.74

Activations Density 0.286%

references to entities, roles, and attributes related to membership and identity

following the articles "a", "an", and sometimes "the"

articles followed by specific words

No Comments

No Known Activations

references to entities, roles, and attributes related to membership and identity

following the articles "a", "an", and sometimes "the"

articles followed by specific words

No Comments

No Known Activations