INDEX

Explanations

requests and conditional phrases indicating desire or preference

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

who have

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

 Савезне

-0.63

 betweenstory

-0.60

 jScrollPane

-0.56

цездатний

-0.55

principalTable

-0.54

 Roskov

-0.54

parsedMessage

-0.53

principalColumn

-0.49

MigrationBuilder

-0.49

 ویکی‌پدیای

-0.47

POSITIVE LOGITS

RegressionTest

0.40

 ligiloj

0.40

 Alegre

0.39

 AssemblyProduct

0.39

 rêver

0.37

chiali

0.37

 schedule

0.36

исленность

0.35

sunter

0.35

Tikang

0.35

Activations Density 0.015%

requests and conditional phrases indicating desire or preference

who have

No Comments

No Known Activations