INDEX

Explanations

technical terms related to data handling and processing

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

<start_of_turn>user marker

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

persistent

-0.46

 continual

-0.40

de

-0.40

te

-0.38

op

-0.36

 ditentukan

-0.36

 Selling

-0.36

te

-0.36

ựng

-0.35

or

-0.35

POSITIVE LOGITS

Hentet

0.82

Diwedd

0.79

 autorytatywna

0.68

devamını

0.66

expandindo

0.65

IsMutable

0.64

 NSCoder

0.63

 localObject

0.63

Obrázky

0.62

 informée

0.61

Activations Density 0.000%

technical terms related to data handling and processing

<start_of_turn>user marker

No Comments

No Known Activations