INDEX

Explanations

prepositions indicating a specific location or place

oai_token-act-pair · gpt-3.5-turbo

instances of the word "in."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

olina

-0.82

keye

-0.78

µ

-0.77

isbury

-0.75

voy

-0.74

letes

-0.73

animal

-0.73

 Olympus

-0.72

burgh

-0.70

riott

-0.69

POSITIVE LOGITS

\<

0.70

 redistributed

0.68

orgetown

0.66

 bargain

0.66

 exempted

0.64

matter

0.62

 exemptions

0.60

wee

0.60

vernment

0.60

/**

0.59

Activations Density 0.000%

prepositions indicating a specific location or place

instances of the word "in."

No Comments

No Known Activations