INDEX

Explanations

phrases indicating conditional or logical relationships within complex scenarios

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

result of, contained in, distributed among

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

 StringTokenizer

-0.56

 noDo

-0.56

исленность

-0.51

 XCTest

-0.49

 cheminée

-0.47

BeginInit

-0.46

ariConfig

-0.45

HtmlAttribute

-0.45

scriptcase

-0.44

 насељу

-0.44

POSITIVE LOGITS

 lateinit

0.47

endregion

0.37

ThroughAttribute

0.36

 futuros

0.35

lateinit

0.33

orElse

0.33

Hochspringen

0.32

OCCURRED

0.31

 solely

0.31

GraphicsUnit

0.31

Activations Density 0.207%

phrases indicating conditional or logical relationships within complex scenarios

result of, contained in, distributed among

No Comments

No Known Activations