INDEX

Explanations

patterns related to mathematical proofs and logical reasoning

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.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

|()

-0.36

 organi

-0.35

-0.34

ConstraintMaker

-0.34

 Torna

-0.34

mancy

-0.33

 uſed

-0.33

wahati

-0.33

*/,

-0.32

 útiles

-0.31

POSITIVE LOGITS

intur

0.48

 invokingState

0.48

ungsbedingungen

0.45

 podjela

0.45

gië

0.45

puente

0.45

AndEndTag

0.44

uevos

0.44

 gyhoeddwyd

0.44

+#+#

0.43

Activations Density 0.277%

patterns related to mathematical proofs and logical reasoning

No Comments

No Known Activations