INDEX

Explanations

phrases related to numerical ranges or durations, especially with hyphens

oai_token-act-pair · gpt-3.5-turbo

phrases indicating the duration of time related to events or processes

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Glas

-0.71

 heights

-0.68

 lows

-0.67

 shields

-0.65

 mirrors

-0.64

Dos

-0.64

 accents

-0.64

 indo

-0.63

RU

-0.63

 Gavin

-0.63

POSITIVE LOGITS

year

1.50

month

1.48

minute

1.44

mile

1.41

member

1.36

hour

1.35

million

1.31

acre

1.30

strong

1.28

nation

1.24

Activations Density 0.080%

phrases related to numerical ranges or durations, especially with hyphens

phrases indicating the duration of time related to events or processes

No Comments

No Known Activations