INDEX

Explanations

the word "and" in various contexts

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

and followed by names and entities

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

SequentialGroup

-0.88

 EconPapers

-0.77

MemoryWarning

-0.72

 termica

-0.62

 '\\;'

-0.57

 оригіналу

-0.54

 térmica

-0.54



-0.54

 initComponents

-0.54

 للاسماء

-0.54

POSITIVE LOGITS

Jîn

0.37

 محفوظة

0.32

Dr

0.31

ViewChild

0.29

 ment

0.28

DeleteBehavior

0.28

 shade

0.27

owohl

0.27

Pil

0.27

诰

0.27

Activations Density 0.071%

the word "and" in various contexts

and followed by names and entities

No Comments

No Known Activations