INDEX

Explanations

dates and events

oai_token-act-pair · gpt-3.5-turbo

punctuation and sentence-ending indicators

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 inconsist

-0.94

 neglig

-0.90

 involuntary

-0.82

neb

-0.82

 intangible

-0.81

 artificially

-0.80

 recycling

-0.79

 manipulating

-0.79

 subconscious

-0.79

 polymorph

-0.79

POSITIVE LOGITS

Tickets

1.61

 Tickets

1.59

 Expect

1.40

 Tune

1.33

 Doors

1.29

 Dates

1.29

 Admission

1.29

 Featuring

1.26

 Tomorrow

1.23

 Join

1.21

Activations Density 0.352%

dates and events

punctuation and sentence-ending indicators

No Comments

No Known Activations