INDEX

Explanations

phrases emphasizing or relating to insights and understanding in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

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

Configuration

google/gemma-scope-9b-pt-res/layer_2/width_16k/average_l0_67

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.2.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

rungsseite

-0.56

 Kimmel

-0.44

 Rendez

-0.42

Stella

-0.41

 cooperate

-0.40

 besta

-0.40

Ladies

-0.40

 "..\..\

-0.39

Wee

-0.39

 gratuitous

-0.39

POSITIVE LOGITS

 isolation

0.69

 insight

0.67

 circumstances

0.63

 Isolation

0.61

Isolation

0.59

 derived

0.59

isolation

0.56

ChildScrollView

0.56

Insight

0.55

 Insight

0.54

Activations Density 0.165%

phrases emphasizing or relating to insights and understanding in various contexts

No Comments

No Known Activations