INDEX

Explanations

references to knowledge 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 @ 17-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_17/width_16k/average_l0_73

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.17.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 useRef

-0.41

enumi

-0.40

 Ferrer

-0.38

 demur

-0.38

 elliptic

-0.38

Dabei

-0.37

 doubly

-0.37

Étienne

-0.37

fastjson

-0.36

def

-0.36

POSITIVE LOGITS

 knowledge

1.48

Knowledge

1.46

 Knowledge

1.46

knowledge

1.40

 KNOWLEDGE

1.32

 pengetahuan

1.13

 conocimiento

1.13

 conhecimento

1.09

 knowled

1.05

NOWLEDGE

1.04

Activations Density 0.053%

references to knowledge in various contexts

No Comments

No Known Activations