INDEX

Explanations

references to event-related activities or services

oai_token-act-pair · gpt-3.5-turbo

references to events

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 hardness

-0.66

 competitiveness

-0.63

anked

-0.62

 mush

-0.61

 suff

-0.61

 tender

-0.60

ni

-0.60

abl

-0.59

 laud

-0.59

âĻ¦

-0.58

POSITIVE LOGITS

 Event

3.46

 Events

2.86

Event

2.81

event

2.60

 EVENT

2.46

 event

2.36

events

2.23

Events

2.19

 events

1.92

 EVENTS

1.27

Activations Density 0.007%

references to event-related activities or services

references to events

No Comments

No Known Activations