INDEX

Explanations

repeated mentions of inputs and outputs in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 0-gemmascope-mlp-65k

Configuration

google/gemma-scope-2b-pt-mlp/layer_0/width_65k/average_l0_72

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

Data Type

float32

Hook Name

blocks.0.hook_mlp_out

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Outputs

-1.02

 outputs

-0.99

 AssemblyCulture

-0.98

parsedMessage

-0.96

AddHtmlAttribute

-0.96

 inputs

-0.94

 BrowserModule

-0.90

rrggbb

-0.90

IsContent

-0.88

 IConfiguration

-0.87

POSITIVE LOGITS

0.61

the

0.61

 only

0.56

 number

0.56

арь

0.54

 head

0.54

delning

0.52

也就是说

0.50

UDIO

0.49

no

0.49

Activations Density 0.017%

repeated mentions of inputs and outputs in various contexts

No Comments

No Known Activations