INDEX

Explanations

phrases related to personal interactions or events

oai_token-act-pair · gpt-3.5-turbo

phrases indicating a significant or unique opportunity

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

 shade

-0.83

 bleed

-0.77

 curb

-0.76

 indirect

-0.76

Sig

-0.75

 buoy

-0.74

 chant

-0.74

 direct

-0.73

 cause

-0.73

fri

-0.73

POSITIVE LOGITS

the

1.53

your

1.52

their

1.50

his

1.40

one

1.36

them

1.34

nothing

1.34

many

1.34

repl

1.33

three

1.33

Activations Density 0.047%

phrases related to personal interactions or events

phrases indicating a significant or unique opportunity

No Comments

No Known Activations