INDEX

Explanations

phrases related to incidents involving law enforcement

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

Data Type

torch.float32

Hook Point

blocks.8.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

 sustainability

-0.90

 merits

-0.88

 merit

-0.87

 innov

-0.84

 democracies

-0.83

 decentral

-0.83

 democracy

-0.80

 unification

-0.80

 monopol

-0.78

 unite

-0.78

POSITIVE LOGITS

Then

1.29

Eventually

1.27

Suddenly

1.23

then

1.13

Later

1.09

Initially

1.09

Luckily

1.04

 Afterwards

1.03

sequently

1.02

 Eventually

1.01

Activations Density 0.257%

phrases related to incidents involving law enforcement

No Comments

No Known Activations