INDEX

Explanations

terms related to scientific concepts, findings, and technical processes

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

past or present actions and outcomes

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

 Мексичка

-0.57

istoitu

-0.56

+#+#

-0.50

〗

-0.50

 himo

-0.49

parsedMessage

-0.49

enumi

-0.49

PyExc

-0.48

NEC

-0.47

DoubleQuotes

-0.47

POSITIVE LOGITS

InstrumentedTest

0.53

 dezelve

0.53

 zelve

0.48

ówczas

0.47

-------

0.47

 operações

0.47

 assistência

0.47

 typelib

0.47

 zoude

0.46

 medesimo

0.46

Activations Density 0.893%

terms related to scientific concepts, findings, and technical processes

past or present actions and outcomes

No Comments

No Known Activations