INDEX

Explanations

elements related to programming terminology and library/package references

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

references like [@word]

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-65k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_65k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

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

 purpoſe

-0.61

 raiſ

-0.60

 ſever

-0.58

 whoſe

-0.58

 fubject

-0.57

 uſed

-0.57

 juſ

-0.56

 leaſt

-0.56

 pleaſure

-0.55

/*

-0.55

POSITIVE LOGITS

1.47

'@

1.02

(@

0.96

at

0.95

"@

0.93

"@

0.93

/@

0.90

="@

0.90

=@

0.90

Activations Density 0.183%

elements related to programming terminology and library/package references

references like [@word]

No Comments

No Known Activations