INDEX

Explanations

conjunctions and indicators of relationships among multiple clauses or concepts

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

Preceding words of "and"

np_acts-logits-general · gemini-2.0-flash

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 31-gemmascope-res-16k

Configuration

google/gemma-scope-9b-it-res/layer_31/width_16k/average_l0_76

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.31.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 réalis

-0.50

WindowConstants

-0.49

Personensuche

-0.47

 apó

-0.45

 considér

-0.41

 élève

-0.39

 ainfi

-0.37

Trailing

-0.36

 gouvernements

-0.36

 Comercio

-0.35

POSITIVE LOGITS

 taken

1.25

 inspected

1.20

 analyzed

1.20

 thrown

1.18

 eaten

1.18

 examined

1.17

 evaluated

1.16

 treated

1.15

 reviewed

1.14

 brought

1.13

Activations Density 0.619%

conjunctions and indicators of relationships among multiple clauses or concepts

Preceding words of "and"

No Comments

No Known Activations