INDEX

Explanations

words related to code generation

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

generated by

np_acts-logits-general · gemini-2.5-flash-lite

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

GraphicsUnit

-0.71

tonode

-0.62

PerformLayout

-0.62

 BorderSide

-0.52

OGND

-0.52

 Theſe

-0.50

 houſe

-0.50

 moeite

-0.48

 ſtre

-0.48

oredCriteria

-0.48

POSITIVE LOGITS

 MyBatis

0.45

CodeAttribute

0.38

 MainAxisSize

0.37

GeneratedCode

0.33

 незавершена

0.33

TokenNameLBRACE

0.32

__(/*!

0.32

sproz

0.32

龚

0.31

 нрави

0.30

Activations Density 0.000%

words related to code generation

generated by

No Comments

No Known Activations