INDEX

Explanations

punctuation and file format indicators related to programming

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

code imports and library functions

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

parsedMessage

-0.93

AndEndTag

-0.83

 '\\;'

-0.80

 disambiguazione

-0.80

Tikang

-0.78

 ujednoznacz

-0.74

 незавершена

-0.73

 fieldNum

-0.72

 ProtoMessage

-0.71

 queſta

-0.71

POSITIVE LOGITS

Slf

0.49

cassert

0.46

0.41

cstdio

0.30

 errors

0.30

data

0.29

).

0.29

error

0.28

 Pública

0.28

Activations Density 0.001%

punctuation and file format indicators related to programming

code imports and library functions

No Comments

No Known Activations

punctuation and file format indicators related to programming

code imports and library functions

No Comments

No Known Activations