INDEX

Explanations

phrases related to events happening for the first time in a particular context or history

oai_token-act-pair · gpt-3.5-turbo

phrases indicating significant first-time events or milestones

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

asty

-0.73

phen

-0.70

 tomat

-0.64

ophe

-0.63

urous

-0.63

oll

-0.62

ancies

-0.62

krit

-0.61

MIC

-0.61

haar

-0.60

POSITIVE LOGITS

ever

0.99

 ever

0.98

 EVER

0.93

 since

0.85

 imaginable

0.81

 anniversary

0.68

 Scotia

0.67

Ever

0.65

 practicable

0.63

 consecut

0.62

Activations Density 0.041%

phrases related to events happening for the first time in a particular context or history

phrases indicating significant first-time events or milestones

No Comments

No Known Activations