INDEX

Explanations

programming constructs and elements in code

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

code identifiers

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

 nakalista

-0.41

 larmes

-0.41

 secours

-0.38

 blessé

-0.37

 épaules

-0.37

 genoux

-0.36

 souligne

-0.36

 orígenes

-0.35

 Absicht

-0.35

 samfun

-0.35

POSITIVE LOGITS

interopRequire

0.60

didSet

0.54

ugeot

0.52

 pedest

0.52

 bomb

0.52

 AssemblyProduct

0.52

 EconPapers

0.52

MemoryWarning

0.52

 penit

0.51

 Infórmanos

0.51

Activations Density 0.032%

programming constructs and elements in code

code identifiers

No Comments

No Known Activations