INDEX

Explanations

terms related to dependencies in various contexts

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

words ending in -ency

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

“

-0.50

“

-0.49

WebServlet

-0.47

-0.46

Rot

-0.44

<blockquote>

-0.44

frac

-0.41

-0.40

multicolumn

-0.40

“(

-0.40

POSITIVE LOGITS

ency

1.19

gency

1.01

ENCY

0.96

ancy

0.94

urgency

0.92

iency

0.92

idency

0.90

encies

0.90

sistency

0.87

 kasarigan

0.85

Activations Density 0.018%

terms related to dependencies in various contexts

words ending in -ency

No Comments

No Known Activations