INDEX

Explanations

references to specific individuals, particularly related to political figures

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

akens

-0.70

æ©

-0.69

bledon

-0.68

 transact

-0.67

published

-0.66

 rehabilit

-0.64

 lender

-0.64

controlled

-0.62

 startup

-0.62

 bidder

-0.62

POSITIVE LOGITS

 Alright

1.33

Hmm

1.24

 Nope

1.18

 Yeah

1.12

 Seems

1.10

 Looks

1.08

 Guys

1.07

Wow

1.07

 Sorry

1.07

 Sounds

1.05

Activations Density 0.058%

references to specific individuals, particularly related to political figures

No Comments

No Known Activations