INDEX

Explanations

phrases that emphasize improvement or positive outcomes

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

control flow constructs

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

fjspx

-0.52

RegressionTest

-0.50

-0.43

blockList

-0.41

 verbe

-0.41

참고

-0.41

addPreferredGap

-0.40

VYMaps

-0.40

évaluateur

-0.40

存于互联网档案馆

-0.39

POSITIVE LOGITS

uxxxx

0.39

 lovely

0.38

getragen

0.37

󠁮

0.34

AutoScaleMode

0.34

multirow

0.34

principalTable

0.34

 titolata

0.33

]]);

0.33

setCellValue

0.33

Activations Density 0.022%

phrases that emphasize improvement or positive outcomes

control flow constructs

No Comments

No Known Activations