INDEX

Explanations

references to structured data or states in a programming context

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

snapshot

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

odils

-0.47

Tikang

-0.46

isdir

-0.44

RSpec

-0.40

expandindo

-0.38

<?

-0.37

">//

-0.36

락

-0.35

RunWith

-0.35

 BoxFit

-0.35

POSITIVE LOGITS

Snapshot

1.18

snapshot

0.80

 Snapshot

0.78

 snapshot

0.78

Snapshots

0.68

StoryboardSegue

0.66

 snapshots

0.64

DataSnapshot

0.59

nestjs

0.53

Sna

0.51

Activations Density 0.000%

references to structured data or states in a programming context

snapshot

No Comments

No Known Activations