INDEX

Explanations

The neuron seems to identify phrases related to projects, initiatives, and plans

oai_token-act-pair · gpt-3.5-turbo

declarative statements about a project's significance or function

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

fire

-0.69

esm

-0.66

bury

-0.64

 resorts

-0.64

isms

-0.63

sburg

-0.63

truth

-0.62

 loses

-0.62

ville

-0.62

gom

-0.61

POSITIVE LOGITS

 slated

1.21

 expected

1.20

 scheduled

1.06

 intended

1.03

 reminiscent

0.99

 aimed

0.99

 currently

0.97

 meant

0.93

expected

0.93

 believed

0.93

Activations Density 0.221%

The neuron seems to identify phrases related to projects, initiatives, and plans

declarative statements about a project's significance or function

No Comments

No Known Activations