INDEX

Explanations

specific numerical codes and notations relevant to structure and organization in various contexts

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

closing bracket, parenthesis, or semicolon

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

GroupLayout

-0.62

الدراسه

-0.60

躇

-0.57

 queſto

-0.57

enumii

-0.57

㤵

-0.56

-0.55

纐

-0.54

iſche

-0.54

addPreferredGap

-0.53

POSITIVE LOGITS

penup

0.42

 мәкал

0.34

EndGlobalSection

0.34

Anyways

0.33

 hidrá

0.31

новниш

0.31

 Anyways

0.29

ConfigureAwait

0.29

 genoemd

0.28

Crossref

0.28

Activations Density 0.002%

specific numerical codes and notations relevant to structure and organization in various contexts

closing bracket, parenthesis, or semicolon

No Comments

No Known Activations