INDEX

Explanations

instances of the word "say" and its variations in different contexts

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

introduce direct speech

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

 OnEnable

-0.49

sidemargin

-0.49

initializeApp

-0.46

 Mani

-0.45

blogspot

-0.44

 pools

-0.43

 GenerationType

-0.41

PullParser

-0.41

 pool

-0.40

┻

-0.40

POSITIVE LOGITS

say

0.68

say

0.67

SAY

0.66

Say

0.62

 kysy

0.60

Say

0.59

saying

0.58

Saying

0.58

 sayin

0.58

 Saying

0.58

Activations Density 0.028%

instances of the word "say" and its variations in different contexts

introduce direct speech

No Comments

No Known Activations