INDEX

Explanations

phrases that express the concept of getting rid of something

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 1-gemmascope-mlp-16k

Configuration

google/gemma-scope-2b-pt-mlp/layer_1/width_16k/average_l0_105

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.1.hook_mlp_out

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

rid

-0.62

]<<

-0.62

rungsseite

-0.62

 Accesat

-0.58

DebuggerNonUser

-0.58

érables

-0.57

Legături

-0.57

ljeno

-0.56

BeginInit

-0.52

putAll

-0.52

POSITIVE LOGITS

 autorytatywna

0.65

 propOrder

0.64

SourceChecksum

0.61

 ghiaccio

0.60

principalColumn

0.59

 Chwiliwch

0.57

 figliu

0.52

 proprietà

0.49

 Padma

0.49

 vestito

0.48

Activations Density 1.798%

phrases that express the concept of getting rid of something

No Comments

No Known Activations