INDEX

Explanations

phrases related to known concepts or established information

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

known to

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

AnimationsModule

-0.51

 dié

-0.43

 spørs

-0.41

 méri

-0.40

 propOrder

-0.39

 Reſ

-0.37

Perspective

-0.37

länger

-0.36

setVerticalGroup

-0.35

Dimension

-0.34

POSITIVE LOGITS

 Known

0.93

 known

0.93

 KNOWN

0.91

Known

0.85

known

0.84

 notorious

0.84

 notoriously

0.82

KNOWN

0.82

 notori

0.75

 know

0.66

Activations Density 0.038%

phrases related to known concepts or established information

known to

No Comments

No Known Activations