INDEX

Explanations

concepts related to oversight and evaluation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 21-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_21/width_16k/average_l0_129

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.21.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

 itself

-0.78

itself

-0.64

its

-0.56

 яке

-0.54

 vœux

-0.52

 Itself

-0.51

Its

-0.50

 которое

-0.49

 enfans

-0.49

Its

-0.48

POSITIVE LOGITS

 themselves

0.94

themselves

0.82

 amelyek

0.75

 jotka

0.75

 которые

0.70

 diejenigen

0.67

 neler

0.66

 eivät

0.65

 abstractions

0.64

 които

0.63

Activations Density 3.839%

concepts related to oversight and evaluation

No Comments

No Known Activations