INDEX

Explanations

mentions of specific locations and organizations, along with a focus on issues and actions related to those entities

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs768-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

768

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

 manif

-0.87

wagen

-0.71

 bearer

-0.67

 wholes

-0.66

 itself

-0.66

 pellets

-0.65

 whichever

-0.65

 cassette

-0.64

 machine

-0.61

 champ

-0.61

POSITIVE LOGITS

 SHARES

0.96

 Advertisement

0.94

 Updated

0.90

"]=>

0.89

ANT

0.86

Reporting

0.85

MORE

0.84

Updated

0.84

Update

0.84

Written

0.84

Activations Density 2.255%

mentions of specific locations and organizations, along with a focus on issues and actions related to those entities

No Comments

No Known Activations