INDEX

Explanations

expressions of personal opinions or hypothetical statements

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 40-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_40/width_16k/average_l0_125

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.40.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

دانشنامهٔ

-0.54

 autorytatywna

-0.53

webElementXpaths

-0.52

IsMutable

-0.47

Hentet

-0.47

 Italijanski

-0.46

 оригіналу

-0.46

UrlResolution

-0.46

addGap

-0.46

Chham

-0.45

POSITIVE LOGITS

say

3.20

say

2.73

 saying

2.59

Say

2.53

Say

2.44

SAY

2.41

 said

2.36

saying

2.22

 says

2.22

SAY

2.13

Activations Density 0.935%

expressions of personal opinions or hypothetical statements

No Known Activations