INDEX

Explanations

programming-related terms and function calls

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

handling code input

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

</tfoot>

-0.45

η

-0.37

tgz

-0.37

ftagPool

-0.36

cussed

-0.36

 splits

-0.35

 brave

-0.35

andin

-0.35

 gorge

-0.35

arity

-0.34

POSITIVE LOGITS

 CURIAM

0.68

parsedMessage

0.61

RTHOOK

0.59

/*++

0.59

dafx

0.57

$_['

0.57

contentLoaded

0.57

windows

0.54

stdafx

0.54

hObject

0.54

Activations Density 0.043%

programming-related terms and function calls

handling code input

No Comments

No Known Activations