INDEX

Explanations

corporate or organizational names and related domains

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

domain names, file paths, and identifiers

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

 disambiguazione

-0.61

 ujednoznacz

-0.61

utilisons

-0.45

 Chwiliwch

-0.45

complexContent

-0.43

География

-0.40

COUVER

-0.39

artifactId

-0.37

labus

-0.36

InputTagHelper

-0.35

POSITIVE LOGITS

GEBURTS

0.59

 oprot

0.50

 विश्वसनीयता

0.48

onavirus

0.41

ournal

0.38

 VizieR

0.38

enderror

0.38

 Walkover

0.36

Fonti

0.36

ferrer

0.36

Activations Density 0.005%

corporate or organizational names and related domains

domain names, file paths, and identifiers

No Comments

No Known Activations