INDEX

Explanations

signs of equality or equivalence in code, particularly in programming language syntax

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

händler

-0.48

 którą

-0.46

riever

-0.46

IContainer

-0.43

Brien

-0.42

antren

-0.42

dataProvider

-0.42

 keempat

-0.42

 която

-0.42

setcounter

-0.41

POSITIVE LOGITS

==

0.66

 =====

0.65

('=

0.61

'=

0.60

"=

0.60

Diweddarwch

0.59

windowFixed

0.57

=====

0.57

("="

0.56

 raiſ

0.56

Activations Density 0.043%

signs of equality or equivalence in code, particularly in programming language syntax

No Comments

No Known Activations