INDEX

Explanations

words related to management or leadership

oai_token-act-pair · gpt-3.5-turbo

phrases indicating relationships or connections between people or concepts

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

ãĤ´ãĥ³

-0.84

Released

-0.75

 lamented

-0.74

BuyableInstoreAndOnline

-0.72

 penned

-0.72

 Released

-0.71

 Â®

-0.64

âī

-0.63

wang

-0.63

 Though

-0.63

POSITIVE LOGITS

 everybody

1.55

 somebody

1.53

1.36

 anybody

1.35

['

1.27

 stuff

1.14

 âĢ¦"

1.14

 guys

1.05

us

1.04

 whoever

1.03

Activations Density 0.797%

words related to management or leadership

phrases indicating relationships or connections between people or concepts

No Comments

No Known Activations