INDEX

Explanations

words related to dates, times, and events

oai_token-act-pair · gpt-3.5-turbo

titles and names of artistic works

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 policymakers

-0.81

.*

-0.77

EA

-0.76

 broadly

-0.76

GPU

-0.76

 healthcare

-0.75

 effectively

-0.74

 potentially

-0.73

 sanctions

-0.73

 regulatory

-0.72

POSITIVE LOGITS

 Goodbye

1.42

 Trouble

1.40

 Forgotten

1.32

 Lover

1.31

 Wedding

1.30

 Falling

1.29

 Haunted

1.27

 Corpse

1.25

 Wicked

1.24

 Ruin

1.24

Activations Density 0.310%

words related to dates, times, and events

titles and names of artistic works

No Comments

No Known Activations