INDEX

Explanations

recurring or repetitive terms, particularly those starting with "rec."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 27-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_27/width_16k/average_l0_118

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.27.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

AnimationsModule

-0.44

поте

-0.42

 humer

-0.41

 initComponents

-0.41

Far

-0.41

Far

-0.41

atiche

-0.41

setBold

-0.39

 Sili

-0.39

 pous

-0.38

POSITIVE LOGITS

REC

0.88

Rec

0.84

rec

0.83

rec

0.76

REC

0.74

Rec

0.72

protoimpl

0.67

Rek

0.61

recs

0.60

rek

0.59

Activations Density 0.177%

recurring or repetitive terms, particularly those starting with "rec."

No Comments

No Known Activations