INDEX

Explanations

proper nouns, particularly names and terms related to "Rome" or "Roman."

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

Tokens starting with "Nom", "Rom", "Om", or "Gom"

np_acts-logits-general · gemini-2.0-flash

Nom, Rom, Gom, Om prefixes

np_acts-logits-general · gemini-2.5-flash-lite

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

GenerationType

-0.46

ValueStyle

-0.44

 evaluación

-0.37

 dicho

-0.36

 médicas

-0.36

 respiración

-0.36

BASELINE

-0.35

 metodología

-0.35

 espirituales

-0.35

afterEach

-0.35

POSITIVE LOGITS

Gom

0.59

OMS

0.57

gom

0.56

oms

0.55

lom

0.53

gom

0.52

Lom

0.52

Gom

0.52

)_/¯

0.52

omi

0.51

Activations Density 0.275%

proper nouns, particularly names and terms related to "Rome" or "Roman."

Tokens starting with "Nom", "Rom", "Om", or "Gom"

Nom, Rom, Gom, Om prefixes

No Comments

No Known Activations