INDEX

Explanations

personal details and relationships within a narrative.

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

possessive pronouns, particularly "his" and "her", often in contexts related to personal actions, belongings, or attributes.

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

possessive pronouns and references to personal ownership.

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

Possessive pronouns followed by related nouns

np_acts-logits-general · gemini-2.0-flash

possessive pronouns followed by possessive nouns

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ValueStyle

-0.73

Portale

-0.65

InjectAttribute

-0.65

Personensuche

-0.62

脚注の使い方

-0.62

quelles

-0.61

ValueGeneration

-0.59

 ExecuteAsync

-0.58

ContentAsync

-0.57

Brainz

-0.57

POSITIVE LOGITS

own

1.18

Own

0.85

Own

0.77

own

0.76

 favorite

0.75

 propia

0.73

 egen

0.71

 própria

0.69

 соб