INDEX

Explanations

references to popular culture and iconic characters

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

its own, favorite Transformer, Ram was

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

 stratégique

-0.56

kään

-0.50

 dentaire

-0.47

 communautaire

-0.46

 éditoriale

-0.44

argout

-0.44

 Italij

-0.43

 industriels

-0.42

 contemporaine

-0.42

 yrity

-0.42

POSITIVE LOGITS

tagHelperRunner

0.60

 NgModule

0.49

ArgsConstructor

0.43

 springfox

0.41

NgModule

0.41

 SqlParameter

0.38

懐かしい

0.38

 costume

0.38



0.38

ftagPool

0.37

Activations Density 0.016%

references to popular culture and iconic characters

its own, favorite Transformer, Ram was

No Comments

No Known Activations