INDEX

Explanations

expressions of personal introspection and motivation.

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

phrases related to personal reflection, sharing experiences, and helping others.

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

phrases or constructions that express personal opinions, thoughts, or statements, often in the first person (e.g. "I think", "Yes, I am", "I find", etc.).

oai_token-act-pair · claude-3-5-sonnet-20240620 Triggered by @bot

Follows pronouns or indefinite articles

np_acts-logits-general · gemini-2.0-flash

contralateral hemisphere, conditions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

 neither

-1.96

 não

-1.88

not

-1.76

 none

-1.74

neither

-1.72

 doesn

-1.72

 tidak

-1.70

 geen

-1.66

 nobody

-1.63

 nothing

-1.63

POSITIVE LOGITS

rungsseite

0.88

 препратки

0.76

 متعلقه

0.74

Autoritní

0.70

 autorytatywna

0.65

 nakalista

0.62

routeProvider

0.61

 مرئيه

0.59

idescreen

0.56

aarrggbb