INDEX

Explanations

descriptions of straightness and alignment in various contexts

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

straight and flat

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

Життєпис

-0.59

Tikang

-0.52

 ब्रेकडाउन

-0.50

MethodManager

-0.50

localctx

-0.48

 otomatig

-0.47

 beginnetje

-0.46

MemoryWarning

-0.44

delay

-0.44

 useSelector

-0.43

POSITIVE LOGITS

 Straight

0.72

Straight

0.70

 straight

0.69

straight

0.66

 straighten

0.54

IntoConstraints

0.54

 Flat

0.50

 upright

0.50

Flat

0.50

 straightening

0.49

Activations Density 0.062%

descriptions of straightness and alignment in various contexts

straight and flat

No Comments

No Known Activations