INDEX

Explanations

terms related to projections and projectors

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

 projection

-1.25

 projecting

-0.96

projection

-0.92

 Projection

-0.82

 projections

-0.81

Projection

-0.75

/\.

-0.73

ceased

-0.69

bolista

-0.68

 projected

-0.67

POSITIVE LOGITS

tagext

0.65

tagHelper

0.61

 Energies

0.60

Aq

0.59

Aq

0.59

AutoScaleMode

0.58

[*]

0.58

 Upon

0.57

Pod

0.57

sī

0.55

Activations Density 0.042%

terms related to projections and projectors

No Comments

No Known Activations